Direct current stimulation stem cell culture equipment

Abstract

The invention relates to the field of cell culture devices, in particular to direct current stimulation stem cell culture equipment which comprises an acquisition module, an analysis module and an electrical stimulation module. The acquisition module is used for acquiring microspur image information in a preset culture range, associating acquisition time as a timestamp with the microspur image information, and extracting cell contour features in the microspur image information. According to the scheme, the cell contour and arrangement characteristics are extracted through the acquisition module, the analysis module combines a mapping table and a morphological evolution model, adaptive stimulation instructions are dynamically generated according to the current differentiation stage and morphological characteristics of the cells at different moments and regions, and the electrical stimulation module performs accurate execution. The problem that when stem cells are cultured through an electrical stimulation scheme with fixed parameters, the cell differentiation form and electrical stimulation have remarkable adaptation contradiction is solved, the differentiation efficiency and targeting performance are improved, multi-stage synchronous differentiation, period shortening and multi-type cell adaptation are achieved, and cell activity and differentiation stability are guaranteed.

Description

Technical Field

[0001] This invention relates to the field of cell culture devices, and more specifically to a direct current stimulation stem cell culture device. Background Technology

[0002] Electrical stimulation, as a core regulatory means for the directed differentiation of stem cells, has the advantages of being non-invasive, allowing for precise intervention in cell signaling pathways, and avoiding the toxic side effects of chemical inducers. It has been widely used in the differentiation and culture of various cell types, such as osteoblasts, chondrocytes, and neuron-like cells, providing key technical support for tissue engineering repair and regenerative medicine research.

[0003] The current mainstream stem cell electrostimulation differentiation technology (represented by the scheme disclosed in Chinese patent application CN119405901A) involves applying a pulsed direct current electric field with fixed parameters to stem cells in the culture flask during in vitro stem cell culture. This electric field drives the directional migration of charged particles such as calcium and potassium ions inside and outside the cells, thereby activating signaling pathways related to osteogenic differentiation, such as PI3K / Akt and Wnt, ultimately inducing stem cells to differentiate into the target cell lineage. In practical applications, this electrostimulation differentiation technology not only improves cell survival rate but also enhances directional differentiation efficiency, providing a feasible technical solution for clinical scenarios such as skull defect repair.

[0004] However, the morphological evolution paths of different target cells differ fundamentally, leading to varying dynamic adaptation requirements for electrical stimulation. Electrical stimulation with fixed parameters struggles to match these differences, ultimately resulting in significant adaptation contradictions between cell differentiation morphology and electrical stimulation. For example, when stem cells differentiate into osteoblasts, they gradually transform from a spindle-shaped, dispersed state to a polygonal aggregation, eventually forming mineralized nodules. During this process, the intensity of electrical stimulation needs to be adjusted synchronously with the increase in cell aggregation. When differentiating into chondrocytes, cells first aggregate from dispersed growth into round cell clusters. After the cell clusters stabilize, they then construct cartilage lacunae within them. Only during the lacunae formation stage does the stimulation need to be finely adjusted to avoid disrupting the cell cluster structure. When differentiating into neuron-like cells, cells need to rapidly extend processes and construct synaptic bifurcation. This type of morphological development demands extremely high stimulation precision, requiring low-intensity, high-frequency stimulation for precise guidance to prevent improper stimulation from causing process breakage. If stimulation parameters adapted for osteoblasts are used, they may disrupt unstable cell clusters when applied to chondrocytes, and easily lead to process breakage when applied to neuron-like cells, further exacerbating the adaptation problem for multi-type cell differentiation. Summary of the Invention

[0005] This invention provides a direct current stimulation stem cell culture device to solve the problem of significant mismatch between cell differentiation morphology and electrical stimulation when culturing stem cells with a fixed parameter electrical stimulation scheme.

[0006] To solve the above-mentioned technical problems, this application provides the following technical solution:

[0007] A direct current stimulation (DC) stem cell culture device includes: a data acquisition module for acquiring macroscopic image information within a preset culture range, associating the acquisition time as a timestamp with the macroscopic image information, extracting cell contour features from the macroscopic image information, and identifying and marking arrangement features from the cell contour features; an analysis module for acquiring the acquisition location and acquisition parameters of the macroscopic image information, processing them to obtain the size information of the macroscopic image information, quantifying the cell size information by combining the relative positional relationship between the cell contour features and the macroscopic image information, and obtaining morphological features; storing an initial mapping table between cell differentiation stages and morphological and arrangement features, matching the morphological and arrangement features of cells in the macroscopic image information with the initial mapping table to obtain the positional distribution of cells at each differentiation stage; and an electrical stimulation module including an electrical stimulation component for providing DC stimulation or pulse stimulation to the preset culture range; the analysis module contains... The system stores cell morphology evolution models of cells at different differentiation stages under various stimulation intensities, types, and durations. The analysis module acquires the current differentiation stage of the cells, combines it with a preset target morphology, and matches the stimulation intensity, type, and duration within the cell morphology evolution model. Then, considering the current cell location distribution at each differentiation stage, and aiming for the shortest differentiation time, it processes the data to obtain the stimulation intensity and type corresponding to each region within a preset culture range at different times. The system generates stimulation commands by combining the time with the corresponding region, stimulation intensity, and type, and sends these commands to the electrical stimulation module. Upon receiving the stimulation commands, the electrical stimulation module processes the region, stimulation type, and stimulation intensity into operating parameters for the electrical stimulation component. Based on these operating parameters, it applies current to the electrical stimulation component, including current magnitude, time, and location.

[0008] Furthermore, the morphological features include cell nuclear features, cell protrusion features, cell outline boundary features, and size features. The arrangement features include cell density distribution features, cell arrangement direction features, and key intercellular structural features. The analysis module calculates the two-dimensional length and width of the cell and the total cell area based on the macro image size information and the cell outline boundary coordinates, and uses the two-dimensional length and width of the cell and the total cell area as size features. The cell nuclear outline is segmented to calculate the cell nuclear area, and the nuclear-cytoplasmic ratio is obtained by comparing it with the total cell area. The nuclear-cytoplasmic ratio and the cell nuclear area are used as cell nuclear features. The pixel set of cell protrusion outlines is extracted, and the protrusion length, width, and inflection points are calculated to determine the protrusion shape and size. The number of protrusions is then counted, and the protrusion shape, size, and number of protrusions are used as cell protrusion features. The outline boundary regularity is quantified by the ratio of the actual outline length to the perimeter of the smallest circumscribed polygon, and the outline boundary regularity is used as cell outline boundary features. In the process of acquiring key intercellular structural features, the analysis module combines the identified cell protrusion features and analyzes the extension direction, overlapping area, and connection point information of adjacent cell protrusions through protrusion shape analysis to determine whether key intercellular structures are formed.

[0009] Furthermore, the electrical stimulation component includes parallel-structured electrodes and a DC adapter component. The inner side of the culture flask has a recessed culture area. The electrodes are fixedly connected to the culture area, and the parallel structure of the electrodes forms a flushing channel with the mouth of the culture flask. The DC circuit of the DC adapter component passes through the culture area and is fixedly connected to it. The analysis module uses the direction from the center of the culture range towards the mouth of the culture flask as a reference direction, establishes reference coordinates based on the center of the culture range and the reference direction, and marks the position information of each area within the culture range in the reference coordinates. The stimulation types include pulse stimulation and DC stimulation.

[0010] The electrical stimulation module stores a parameter matching mapping table containing stimulation intensity, stimulation type, and the magnitude of the current and voltage applied to the electrical stimulation component. The module matches the received stimulation intensity and type against this mapping table to determine the electrical stimulation component and the magnitude of the applied current or voltage. It then adjusts the timing of the applied current to the electrical stimulation component based on the stimulation duration, using this timing, magnitude, and voltage as operating parameters. The electrical stimulation module is electrically connected to the electrodes and DC adapter, and is used to control the timing, magnitude, and voltage of the applied current to the electrodes and DC adapter according to the operating parameters.

[0011] Furthermore, a metal sheet is embedded in the side of the culture flask in the culture area. The metal sheet is embedded in the inner wall of the culture flask, and its outer surface is exposed to the air inside the culture flask. The metal sheet is electrically connected to a power supply device through a current regulating device. A fluorescent fiber optic temperature measuring device is provided on the outer side of the culture flask in the area corresponding to the metal sheet. The fluorescent fiber optic temperature measuring device is used to detect the temperature of the area in contact with the metal sheet from the outside of the culture flask. The fluorescent fiber optic temperature measuring device sends the real-time temperature signal to the acquisition module, which converts the temperature signal into temperature data and transmits it to the analysis module in real time.

[0012] Furthermore, the device includes multiple metal sheets distributed in a dot matrix pattern within the culture area. A fluorescent fiber optic thermometer stores a unique identifier for each probe. The fluorescent fiber optic thermometer associates the acquired temperature signal with the probe identifier and sends it to the acquisition module. The acquisition module stores a mapping table between the position information of the metal sheets and the probe identifiers. The acquisition module combines the temperature signals received at the same time with the probe identifiers and their positions in the mapping table to obtain temperature distribution data within the culture area, and then sends the temperature distribution data to the analysis module.

[0013] Furthermore, the system includes multiple DC adapter components, whose DC wires are distributed in a network within the cultivation area. Each DC adapter component has a unique identification number associated with the location information of its DC wire within the cultivation area. The electrical stimulation module provides a constant DC current to only one DC adapter component at a time. The electrical stimulation module matches the DC adapter component to the location of the region in the cultivation area according to the stimulation command, and matches the stimulation intensity to the current magnitude. The electrical stimulation module stores the stimulation intensity and corresponding stimulation range of each electrical stimulation component under different operating parameters. The electrical stimulation module first obtains the identification number of the electrical stimulation component by matching the stimulation type. The electrical stimulation module matches the identification number of the electrical stimulation component with the region and stimulation range, and adjusts the operating parameters of the electrical stimulation component corresponding to the identification number according to the stimulation intensity.

[0014] Furthermore, the analysis module stores a temperature mapping table of differentiation stages, differentiation directions, and culture temperature ranges. Based on the current differentiation stage and the differentiation direction corresponding to the preset target morphology, the culture temperature range is matched in the temperature mapping table as a reference temperature range. The reference temperature range is then combined with the current cell position distribution at each differentiation stage to obtain the position distribution of the reference temperature range as a reference temperature distribution. The reference temperature distribution and temperature distribution data are sent to the electrical stimulation module. The parameter matching mapping table also stores heat generation data of each electromagnetic device under different operating parameters. The electrical stimulation module calculates the total heat generation data in each culture area based on the current operating status of each electromagnetic component. The total heat generation data is compared with the temperature distribution data to obtain the actual heat dissipation data. The reference temperature distribution and temperature distribution data are compared to obtain the temperature adjustment range. The temperature adjustment range, heat generation data, and actual heat dissipation data are combined to adjust the operating parameters of each electrical stimulation component without changing the stimulation range and intensity.

[0015] Furthermore, the analysis module invokes the built-in cell morphology evolution model, inputting accumulated environmental stimulus data and current cell morphology and arrangement characteristics data. The environmental stimulus data includes temperature distribution data, electrical stimulation parameters, and nutrient supply-related data, generating a predicted cell differentiation stage. Based on this predicted result, the analysis module generates a sampling distance adjustment command and transmits it to the sampling module. When the cell arrangement characteristics cannot confirm the predicted differentiation stage, the sampling module shortens the sampling distance for macro images and increases the sampling frequency. When the cell arrangement characteristics match the predicted result, the sampling module maintains a preset conventional distance and frequency for acquiring macro images. When the analysis module sends reference temperature distribution and temperature distribution data to the electrical stimulation module, it generates a temperature adjustment command and sends it to the sampling module. Upon receiving the temperature adjustment command, if the cell arrangement characteristics cannot confirm the predicted differentiation stage, the sampling module shortens the macro image sampling distance and uses the conventional frequency as the sampling frequency. If the cell arrangement characteristics confirm the predicted differentiation stage, the sampling distance for macro images is increased and the sampling frequency is slowed down.

[0016] Furthermore, it also includes a culture medium supply circuit, which is electrically connected to the analysis module; the cell morphology evolution model also includes nutrient consumption rate data, and the analysis module acquires the nutrient consumption rate data when matching the stimulus intensity; when the difference between the reference temperature distribution and the temperature distribution data is greater than a preset temperature adjustment threshold, the area greater than the preset temperature adjustment threshold is taken as the temperature adjustment area, and a temperature adjustment command is generated based on the temperature adjustment area and sent to the culture medium supply circuit and the acquisition module. The culture medium circuit lowers or increases the temperature of the culture medium supplied to the higher temperature area and adjusts the culture medium supply volume; after receiving the temperature adjustment command, the acquisition module increases the acquisition frequency of macro images.

[0017] Furthermore, the culture medium circuit has local nozzle components above the parallel paths of each electrode. These nozzle components are fixedly connected to the inner wall of the culture flask at the upper end of the culture area. The culture medium circuit includes separate paths between each local nozzle component and a control valve that controls the volume of culture medium delivered to the local nozzle component. Each separate path also has a temperature regulating device connected to a separate path. Each temperature regulating device and control valve is electrically connected to the analysis module. The analysis module controls the temperature of the culture medium sprayed to each location in the culture area using the unique identification number of the temperature regulating device and the corresponding position of the local nozzle component. It also controls the volume of culture medium sprayed to each location in the culture area using the control valve number and its corresponding position. The analysis module counts the number of temperature adjustments and the corresponding adjustment time based on the time the temperature regulating device is connected to the current.

[0018] Furthermore, the analysis module designates regions that conform to the preset target morphology at the current differentiation stage as successful culture regions, and the time corresponding to the current differentiation stage conforming to the preset target morphology as successful culture time. It acquires the operating parameters of each electrical stimulation component of the electrical stimulation module within the successful culture time, constructs a dynamic model of electrical stimulation type and intensity distribution at each time within each culture region, acquires temperature distribution data within the successful culture time, constructs a dynamic temperature model, acquires the supply time and location of the culture medium within the successful culture time, and constructs a dynamic supply model. Finally, it acquires the morphological and arrangement characteristics of cells in macro images within the successful culture time and constructs a dynamic morphological model. The electrical stimulation dynamic model, temperature dynamic model, supply dynamic model, and morphological dynamic model are aligned over time to perform a collaborative analysis of the parameter change trends and correlations of each model. Based on the collaborative analysis results, the analysis module generates parameter write-back adjustment instructions to write back and optimize the electrical stimulation type, electrical stimulation time, electrical stimulation intensity, culture temperature range, and nutrient consumption rate data. The optimized parameters are then associated with the location information of the successful culture area within the culture area and stored as a default template in the built-in database. When a new batch of culture is started, the analysis module generates a template call instruction to directly apply the default template to the corresponding culture area of ​​the culture flask.

[0019] Furthermore, when applying the default template in a new batch of culture, the analysis module collects electrical stimulation parameters, temperature distribution data, culture medium supply data, and cell morphology and arrangement characteristics data in real time during the actual culture process. It compares these data with the parameters of each dynamic model corresponding to the default template in real time and calculates the parameter deviation value. When the deviation value exceeds a preset threshold, a real-time fine-tuning instruction is generated to dynamically adjust the electrical stimulation type, culture temperature, and culture medium supply volume of the corresponding region. After the new batch of culture is completed, the analysis module merges the complete data of the actually successfully cultured region with the original default template data, performs multi-model collaborative analysis again, generates a template update instruction, iteratively optimizes the parameters and regional correlations of the default template, and stores the results.

[0020] Furthermore, the electrical stimulation module is equipped with an impedance acquisition circuit. While constructing the electrical stimulation dynamic model, the analysis module generates electrode function switching commands to control the lattice-distributed metal sheet electrodes to alternately switch between current injection electrodes and voltage detection electrodes. During the electrode switching process, the impedance acquisition circuit acquires multiple sets of impedance signals and transmits them to the analysis module. The analysis module calls the electrical impedance tomography algorithm to invert the conductivity distribution of the culture medium and extracts data from the cell aggregation region. The cell aggregation region data is integrated into the morphological dynamic model and subjected to four-dimensional collaborative analysis with the electrical stimulation dynamic model, temperature dynamic model, and supply dynamic model. During parameter write-back optimization, the electrical stimulation intensity gradient and culture medium supply density of the corresponding regions are adjusted in a targeted manner based on the cell aggregation region data.

[0021] The basic principle and beneficial effects of this scheme are as follows: Fixed-parameter electrical stimulation can lead to insufficient stimulation of some cells and excessive stimulation of others due to different differentiation progress within the same culture environment. This scheme, however, uses an analysis module to locate the distribution of cells at each differentiation stage and customizes exclusive stimulation schemes for different regions. For example, higher intensity stimulation is applied to the region where cells in the osteogenic differentiation mineralization stage are located, while lower intensity stimulation is applied to the region where cells in the proliferation stage are located. This achieves synchronous and efficient differentiation of cells at multiple stages and avoids the adaptation conflict of a single stimulation parameter to cells in mixed stages.

[0022] The analysis module generates stimulation commands with the goal of minimizing differentiation time. Combined with pre-stored cell morphology evolution models, it can screen for the optimal stimulation combination for each differentiation stage, avoiding differentiation lag caused by inappropriate parameters in fixed-parameter stimulation. Fixed-parameter electrical stimulation often leads to prolonged osteogenic differentiation cycles and slow development of neuron-like cell processes due to mismatches in stimulation intensity or time. This approach, however, accelerates morphological evolution without damaging cells by dynamically matching stimulation parameters. Furthermore, because the stimulation parameters are precisely matched to the cell state, it reduces cell dysfunction caused by excessive stimulation, further ensuring cell survival and achieving the dual effect of controlling efficient stem cell differentiation and protecting cell activity.

[0023] This solution extracts cell contours and arrangement features through an acquisition module, and combines a mapping table and morphological evolution model with an analysis module to dynamically generate appropriate stimulation instructions for cells at different times and in different regions based on the current differentiation stage and morphological characteristics of the cells. The electrical stimulation module executes these instructions precisely, solving the problem of significant mismatch between cell differentiation morphology and electrical stimulation when culturing stem cells with fixed-parameter electrical stimulation protocols. This improves differentiation efficiency and targeting, enables multi-stage synchronous differentiation, shortens the cell cycle, and adapts to multiple cell types, ensuring cell viability and differentiation stability. Attached Figure Description

[0024] Figure 1 This is a functional module structure diagram of a direct current stimulation stem cell culture device in Example 1;

[0025] Figure 2 This is a structural diagram of a DC wire in the cultivation area;

[0026] Figure 3 This is a structural diagram of the parallel electrode in the culture region;

[0027] Figure 4 This is a structural diagram of the metal sheet lattice in the culture region.

[0028] The attached diagram includes the following labels: 1. Culture flask; 2. Culture area; 3. Metal sheet; 4. Parallel electrode; 5. DC wire. Detailed Implementation

[0029] The following will describe the concept and technical effects of the present invention clearly and completely with reference to embodiments, so as to fully understand the purpose, features and effects of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are all within the scope of protection of the present invention.

[0030] Example 1

[0031] like Figure 1 As shown, a direct current stimulation stem cell culture device includes:

[0032] The acquisition module acquires macro image information within a preset culture range (the preset culture range refers to the overall space within the culture flask 1 used for stem cell culture, within which there are multiple recessed independent culture areas 2 (hereinafter referred to as culture areas 2), each culture area 2 being a physical unit of a preset size (such as 120μm×120μm in the embodiment), constituting a regionalized control unit) (within the cell culture area 2 within the culture flask 1), and associates the acquisition time as a timestamp with the macro image information, extracts cell contour features from the macro image information, and identifies and marks the arrangement features from the cell contour features.

[0033] The core hardware of the acquisition module includes an optical imaging component consisting of a 20x-40x high numerical aperture (numerical aperture ≥ 0.65) infinity-corrected objective lens with a 1x magnification adapter lens group and a 460nm-550nm band 16-level brightness adjustable LED ring light source; an image sensing unit consisting of a back-illuminated sCMOS sensor (resolution ≥ 2048×2048 pixels, pixel size 6.5μm) and a USB3.1 / 10GigE interface data transmission module; and a timing and positioning synchronization unit consisting of a high-precision real-time clock chip with timing accuracy ≤ 1μs and an image positioning sensor with positioning accuracy ≤ 50μm. All components are integrated into a custom aluminum housing and powered by a 12VDC wide voltage power supply module.

[0034] For integration of fluorescence channels (such as GFP / RFP) or impedance imaging and optical multimodal synchronization functions, dual fluorescence (GFP / RFP) and electrical impedance tomography (EIT) multimodal pixel-level synchronization can be achieved on the existing bright-field / phase-contrast imaging platform through modular stacking, without the need for customized optical components and with compatibility with the original system interface. The fluorescence upgrade utilizes an infinity-coaxial fluorescence accessory and a 2-position flip-up filter drawer, paired with 470 / 540nm LEDs and closed-loop brightness control, achieving bright-field / GFP / RFP switching of <200ms. The EIT upgrade retains the 8×8 gold-plated electrode array, achieving impedance acquisition of 30ms / frame via an analog switch matrix and a high-precision constant current source / ADC. A 10MHz clock source and hardware trigger ensure optical and impedance synchronization errors of <1ms. Mechanically, a sliding drawer-type fluorescence component and biocompatible silicone-sealed FPC wiring are used. The software provides a multimodal callback interface, enabling four-dimensional synergy between fluorescence labeling (cell viability), impedance imaging (cell aggregation), and the original optical analysis.

[0035] The analysis module acquires the acquisition location and acquisition parameters of macro image information to obtain the size information of macro image information. It combines the cell contour features with the relative positional relationship of macro image information to quantify the cell size information and obtain morphological features. It stores an initial mapping table between cell differentiation stages and morphological and arrangement features. It matches the morphological and arrangement features of cells in macro image information with the initial mapping table to obtain the positional distribution of cells at each differentiation stage.

[0036] The analysis module calculates the two-dimensional length and width of the cell and the total area of ​​the cell based on the size information of the macro image and the cell contour boundary coordinates, using these dimensions as size features. It then segments the cell nucleus contour to calculate the cell nucleus area, comparing it with the total cell area to obtain the nucleocytoplasmic ratio, using this ratio and area as cell nucleus features. Finally, it extracts the pixel set of cell protrusion contours, calculates the protrusion length, width, and inflection points to determine the protrusion shape and size, and counts the number of protrusions, using their shape, size, and number as cell protrusion features. Finally, it quantifies the contour boundary regularity by the ratio of the actual contour length to the perimeter of the smallest circumscribed polygon, using this regularity as a cell contour boundary feature.

[0037] After capturing macro images, the acquisition module first performs Gaussian filtering for noise reduction and adaptive histogram equalization to enhance edge contrast. Then, it corrects uneven illumination through grayscale normalization. The Otsu algorithm is used to automatically separate the cell foreground and background, combined with the Canny algorithm to locate edges and complete broken segments. For overlapping cells, distance transformation is performed to locate seed points, and the watershed algorithm is used to achieve accurate separation. Subsequently, a contour detection algorithm is used to capture the set of cell edge pixels. False contours are removed by filtering according to typical cell size and roundness. Finally, the perimeter, area, convex hull area, Euler number, and Fourier descriptor are quantified and extracted to form complete cell contour feature data, which is synchronously associated with timestamps and acquisition location information.

[0038] Specifically, Gaussian filtering uses a 5×5 filter kernel, traversing image pixels to complete noise reduction. The filtering formula is as follows:

[0039] (σ=1.0) Filtered coordinates The pixel grayscale value at that location; : The horizontal and vertical coordinates of the image pixels (with the center of the filter kernel as the origin; in this implementation, the center of the filter kernel is the geometric center point of the 5×5 filter kernel (i.e., the pixel position in the 3rd row and 3rd column). : Standard deviation of Gaussian function (controls the smoothness of filtering, here the value is 1.0); Pi (π)

[0040] When performing adaptive histogram equalization, set the grid size to 8×8 and the contrast limit parameter to 2.0 (corresponding to clipLimit=2.0 in the OpenCV function), and adjust the grayscale distribution grid by grid; when performing grayscale normalization, follow the formula: (I represents the original pixel value,) , These are the maximum and minimum gray values ​​of the image, respectively, mapping the gray values ​​to the 0-255 range. This represents the normalized grayscale value of the target pixel. The grayscale value of the pixel to be processed in the original image; The maximum grayscale value of the original image; 255 is the minimum gray value of the original image; 255 is the maximum quantization value of the grayscale image (corresponding to the upper limit of the 8-bit grayscale range).

[0041] During cell segmentation, first calculate the probability distribution of the gray-level histogram. According to the formula ( , Foreground / background pixel ratio , (The grayscale value is the mean of the foreground / background grayscale), and the grayscale value corresponding to the largest inter-class variance is selected as the threshold. The inter-class variance between the foreground and background is used. A low threshold of 50 and a high threshold of 150 are then set, and edges are filled using gradient calculation, non-maximum suppression, and dual-threshold filtering. Finally, a distance transformation is performed on the overlapping regions. ( (where the coordinates are the background pixels), the peak distance is used as the seed point, and segmentation is performed using the watershed algorithm. Wherein, coordinates The distance transformation value at that location (the shortest distance to the nearest background pixel); The pixel coordinates of the foreground (cell) region to be calculated; The coordinates of any pixel in the background area; To extract all background pixels The minimum distance.

[0042] When extracting and quantizing cell contours, a chain-encoding method is first used to traverse the connected regions of the binary image to obtain the contour pixel coordinate set; then, areas of 50-5000 pixels and roundness of 0.2-1.0 are retained (roundness formula). The outline (where A is the area of ​​the outline and L is the perimeter); where This represents the roundness of the cell outline; the closer the value is to 1, the closer the outline is to a perfect circle. The area of ​​the cell outline is obtained by pixel counting. The perimeter of the cell outline is obtained by summing the distances between the coordinates of the outline pixels; then the perimeter is calculated. Area A (pixel count method), convex hull area Euler number E, through Fourier descriptor ( The shape is described by a sequence of contour coordinates (N is the number of contour points).

[0043] in , The total perimeter of the cell outline; For the outline of the first The coordinates of each pixel; For the outline of the first The coordinates of each pixel; This involves summing the distances between all adjacent contour pixels. : No. Fourier descriptor ( ); : The total number of pixels on the cell outline; : Outline of the first Complex coordinates of pixels ( , The x and y coordinates are pixels. (Imaginary unit).

[0044] Morphological features include nuclear features, cell protrusion features, cell outline boundary features, and size features. Arrangement features include cell density distribution features, cell arrangement direction features, and key intercellular structural features. Identification of key intercellular structural features is based on depth-related protrusion information; first, the coordinates of the cell protrusion endpoints are extracted (…). , ), through formula Calculate the endpoints of adjacent cell processes ( , )and( , The physical spacing of the pixels is considered, and when d ≤ 5μm, the key structural features are identified as protruding connections. This is also considered in conjunction with the overlapping pixel area of ​​the contour. With the area of ​​a single cell ,pass Calculate the overlap ratio; if O ≥ 10%, the key structural feature is determined to be cell adhesion. Alternatively, define a continuous region based on cell clusters and count the total number of cells T within that region (based on contour recognition results); if a cell has at least one protrusion endpoint distance from its neighboring cell... Cells ≤5μm are considered to participate in the connection, and their number C is counted; when When the percentage is ≥70%, the cells in the region are determined to be arranged in a network structure, stacked and adhered in a layered structure, and locally connected in a clustered structure.

[0045] An electrical stimulation module includes an electrical stimulation component for providing direct current or pulsed stimulation to a predetermined culture range. The electrical stimulation component includes parallel electrodes and a direct current adapter. Figure 3 As shown, the inner side of the culture flask 1 has a recessed culture area 2. The electrode (one side) is fixedly bonded to the culture area 2 (bottom), and the electrode has a parallel structure (as shown in the image). Figure 3 The parallel electrode 4 (shown) forms a flushing channel with the mouth of the culture flask 1; the DC wire 5 of the DC adapter component passes through the culture area 2 and is fixedly connected to it. The parallel electrode 4 and the DC wire 5 can be made of materials with extremely strong chemical inertness, excellent corrosion resistance, and superior biocompatibility, such as platinum. The analysis module uses the direction from the center of the culture range towards the mouth of the culture flask 1 as a reference direction, establishes a reference coordinate system based on the center of the culture range and the reference direction, and marks the position information of each area within the culture range in the reference coordinate system; the stimulation types include pulse stimulation (generated by the parallel electrode 4) and DC stimulation (generated by the DC adapter component).

[0046] like Figure 4As shown, a metal sheet 3 (which can be made of a chemically inert, corrosion-resistant, and biocompatible material such as platinum) is embedded in the side of the culture flask 1 in the culture area 2. The metal sheet 3 is embedded in the inner wall of the culture flask 1, and its outer surface is exposed to the air inside the culture flask 1. The metal sheet 3 is electrically connected to the power supply via a current regulating device to form an electrode, which is an electrode in the electrostimulation component. A fluorescent fiber optic temperature measuring device is provided on the outer side of the culture flask 1 in the area corresponding to the metal sheet 3. The fluorescent fiber optic temperature measuring device contacts the area on the outer side of the culture flask 1 corresponding to the metal sheet 3, and is used to detect the temperature of the area in contact with the metal sheet 3 from the outside of the culture flask 1. The fluorescent fiber optic temperature measuring device sends the real-time temperature signal to the acquisition module, which converts the temperature signal into temperature data using an AD converter and then transmits it to the analysis module in real time. This includes multiple DC adapter components, such as... Figure 2 As shown, the DC wires 5 of multiple DC adapter components are distributed in a network within the culture area 2, and the unique identification number of each DC adapter component is associated with the location information of the DC wire 5 within the culture area 2; the electrical stimulation module provides constant DC to only one DC adapter component at a time.

[0047] The analysis module stores cell morphology evolution models of cells at different differentiation stages under different stimulation intensities, types, and durations. The module acquires the current differentiation stage of the cells and combines it with a preset target morphology (the desired cell morphology of stem cells as expected by the technician). It matches the stimulation intensity, type, and duration within the cell morphology evolution model, and then, considering the current cell distribution at each differentiation stage, processes the data to obtain the stimulation intensity and type corresponding to each region within the preset culture area at different times, aiming for the shortest differentiation time. It generates a stimulation command by combining the time with the corresponding region, stimulation intensity, and stimulation type, and sends the command to the electrical stimulation module. Upon receiving the stimulation command, the electrical stimulation module processes the region, stimulation type, and stimulation intensity into operating parameters for the electrical stimulation component. Based on these parameters, it applies current to the component, including current magnitude, time, and location. The electrical stimulation module matches a DC adapter to the region in culture area 2 based on its position within the stimulation command, and matches the current magnitude to the stimulation intensity. The electrical stimulation module stores the stimulation intensity and corresponding stimulation range of each electrical stimulation component under different operating parameters. The electrical stimulation module first obtains the identification number of the electrical stimulation component by matching the stimulation type; the electrical stimulation module matches the identification number of the electrical stimulation component with the region range and the stimulation range, and adjusts the operating parameters of the electrical stimulation component corresponding to the identification number according to the stimulation intensity.

[0048] The analysis module constructs a core objective function to quantify the total differentiation time based on the differentiation stage, location distribution, and preset target morphology of cells in each region of culture region 2. The formula is as follows: ,

[0049] Among them, the differentiation time of stage P cells in region r It is directly related to the cell's need to complete the transformation from its current morphology to the target morphology and the differentiation rate corresponding to the stimulation parameters. The correlation formula is as follows: ,

[0050] R: Total differentiation time for all regions and all stages of cell differentiation to complete the target morphology differentiation (core optimization object, to be minimized); R: Set of regions within the preset culture range (e.g., a 100×100 pixel grid divided by density, r is a unique number for a single region); P: Set of complete stages of stem cell differentiation (e.g., the proliferative phase, early differentiation phase, and maturation phase of mesenchymal stem cell differentiation into osteoblasts, p is a number for a single stage). The stimulus intensity at time t for a cell in region r at stage p. Stimulus type Under the influence of this effect, the time consumed in a single stage to complete the differentiation of this stage; : Morphological difference completion amount for P cells in region r stage (calculated by the analysis module by comparing the contour features, nucleocytoplasmic ratio, and protrusion parameters of the current cell morphology with the target morphology, etc., dimensionless; for example, if the area needs to increase by 30%, then...) ); The differentiation rate of this type of cell under the corresponding stimulation parameters (unit: morphological completion amount / hour) is taken from a preset cell morphology evolution model and satisfies the quantitative correlation. ( The basic differentiation coefficient for stage p. (These are the stimulus intensity sensitivity coefficients, which are constant parameters used in model fitting). : The intensity of electrical stimulation applied to region r at time t (unit: V / cm); : The type of electrical stimulation applied to region r at time t (discrete variable, such as DC voltage encoded as 1, pulse voltage encoded as 2, and the type of adaptation to be matched at stage p).

[0051] To avoid cell damage or differentiation failure due to inappropriate stimulation parameters, the calculation process must adhere to the following constraints to ensure that the optimized results balance short differentiation time with high cell viability, and that the stimulation intensity tolerates the constraints: in, : The minimum effective stimulus intensity for cell differentiation (a preset constant, such as 1V / cm; differentiation cannot be initiated below this value). The maximum tolerance strength of p cells in the proliferative phase (preset by technicians in cell morphology evolution models, such as the proliferative phase). V / cm, differentiation and maturation period V / cm, to avoid high-intensity cell damage.

[0052] Morphological matching constraints: , : The actual morphological vector of the p cell in the r phase of region at time t (composed of parameters such as outline area, roundness, and nucleocytoplasmic ratio); : Preset target shape vector; : Morphological matching degree calculation function (using cosine similarity algorithm); Minimum matching threshold (technicians typically preset a constant of 0.9 to ensure that the differentiation results meet expectations).

[0053] Stimulus type adaptation constraints: , The set of adaptive stimulation types for stage P cells (preset by technicians in cell morphology evolution models, such as only pulsed electrical stimulation during the proliferative phase, i.e.) During the differentiation and maturation period, it adapts to both direct current and pulsed current, i.e. ).

[0054] Region parameter consistency constraints: (r and (adjacent areas) The maximum allowable difference in stimulation intensity between adjacent regions (technicians usually preset a constant of 0.5V / cm) to avoid local stimulation intensity mutations that could lead to an imbalance in the cell growth environment.

[0055] After defining the objective function and constraints, a genetic algorithm is used to efficiently find the optimal solution by traversing the combinations of time points, regions, and stimulus parameters. The specific process is as follows: Each optimization unit corresponds to a certain time t, with a preset complete stimulus scheme for all regions within the cultivation range, expressed as: , : Individual stimulation parameter combination at time t (a complete stimulation parameter scheme for all regions within the preset culture range at a certain time t; an individual is a complete stimulation parameter scheme covering all culture regions 2 at a certain time t). To cultivate the total number of regions 2, each region corresponds to one set of parameters for stimulus intensity E and stimulus type S, forming a complete parameter set. Stimulus intensity (Continuous values, unit: V / cm) Real number encoding is used, directly embedding the values ​​into the individual, with the value range strictly following... (e.g., differentiation stage 1 cell) V / cm); Stimulation type (Discrete values, such as DC = 1, pulse = 2) are encoded using integers, and their values ​​are restricted to the set of adaptation types corresponding to the differentiation stage. Inside.

[0056] Based on a preset population size (e.g., 50), randomly assign values ​​to each optimization unit: for each region r, in Internal random generation real numbers, in random selection The integer encoding ensures that all initial individuals satisfy the basic constraints, forming the initial optimized population.

[0057] Construct a fitness function that combines the shortest total differentiation time with constraint compliance to quantify the merits of each parameter combination: Fitness: Fitness value (the larger the value, the better the parameter combination and the closer it is to the optimal target); , where is the total differentiation time corresponding to this parameter combination; : Constraint violation penalty coefficient (technical personnel preset a constant of 100 to amplify the cost of violations); The degree of violation of the four types of constraints ( For strength tolerance constraints, To adapt constraints to stimulus types For shape matching constraints, (For regional parameter consistency constraints). Constraint weights (e.g.) , , , This amplifies the penalty for core constraints such as shape matching and type adaptation; an example of constraint violation calculation: (The portion exceeding the strength limit is accumulated); (When the type is in violation) or 0 (when it is compliant).

[0058] The population is iteratively updated according to the classic genetic algorithm process, gradually selecting the optimal parameter combination. The steps are as follows: Calculate the fitness value of each individual in the population, sort them from high to low fitness, and select the top 50% of high-quality individuals (e.g., the top 25 if the population size is 50) as parent individuals to ensure that high-quality parameter combinations are preserved. Randomly pair up the parent individuals and exchange parameter fragments in a certain region according to a preset crossover probability (e.g., 0.8) to generate offspring individuals. Example: Parent 1 With parent generation 2 The E-parameter of the cross region 1 yields the offspring. The parameters of the offspring individuals are randomly adjusted according to a preset mutation probability (e.g., 0.05) to avoid the algorithm getting trapped in local optima.

[0059] Mutation: in Regenerate real numbers randomly within the inner circle; Mutation: in Re-select integer encoding. For each offspring individual after crossover and mutation, verify all constraints one by one. If any violation exists (such as...), ... Exceeding limits If there is an adaptation error, the non-compliant parameters will be corrected to the compliant range (e.g., ...). V / cm corrected to V / cm), ensuring that all offspring individuals satisfy the business constraints. Repeat the above stepwise selection process for the optimal parameter combination (technical personnel usually preset an upper limit of 100 iterations) until the rate of change of the optimal fitness value of the population for 10 consecutive generations is <0.1%, and the parameter combination corresponding to the optimal individual has no constraint violations. After the iteration converges, select the individual with the highest fitness value in the population, which is the optimal stimulus parameter combination at time t; decompose the corresponding parameters for each region r from this individual. (Stimulus intensity) and (Stimulus type), serving as the execution parameter for each region r at time t. The optimal stimulus parameters obtained by the algorithm are used to divide key time nodes according to the differentiation process (e.g., 0h, 6h, 12h, 24h), establishing a precise mapping relationship between time, region range, stimulus intensity, and stimulus type. For example, at time 6h, the stimulus parameters corresponding to region 1 (high-density, directionally arranged stage 2 cells) are... V / cm, (Pulsed electrical current); The stimulation parameters corresponding to region 5 (medium density, randomly arranged stage 1 cells) are: V / cm, (Pulsed electricity). Finally, all mapping relationships are integrated to generate standardized stimulation instructions.

[0060] The electrical stimulation module stores a parameter matching mapping table for stimulation intensity, stimulation type, and the magnitude of the current and voltage applied to the electrical stimulation component. The module matches the received stimulation intensity and type against this mapping table to determine the appropriate electrical stimulation component and the required current or voltage. It then adjusts the timing of the current application to the electrical stimulation component based on the stimulation duration. These timing, current magnitude, and voltage are used as operating parameters. The electrical stimulation module is electrically connected to the electrodes and DC adapter, and is used to control the timing and magnitude of the current applied to the electrodes and DC adapter according to the operating parameters.

[0061] The electrical stimulation module converts digital signals into analog reference values ​​via a DAC, which are then amplified and output by a power drive circuit. The DC adapter component adapts the electrode load characteristics to compensate for impedance changes. The power drive circuit mainly includes an H-bridge drive circuit (composed of 4 power MOSFETs, supporting bidirectional current control), a constant current source drive circuit (including operational amplifiers, power MOSFETs, and sampling resistors to achieve accurate and stable current output), a PWM power amplifier circuit (achieving efficient power conversion through duty cycle modulation), a push-pull power amplifier circuit (complementary transistors eliminate crossover distortion, ensuring low-distortion output), and a boost / buck drive circuit (adapting to different voltage requirements). Its amplified output is achieved through signal conversion, drive adjustment, and feedback correction. First, the main control unit analyzes the parameters and then outputs a reference voltage / current signal via a DAC. Then, the operational amplifier compares the reference signal with the actual signal fed back by the sampling resistor, and outputs a drive signal to control the power MOSFETs and other devices to adjust the output power. Voltage amplification can be achieved through a linear amplification path or PWM filtering. Current amplification relies on the voltage-controlled constant current source architecture to lock the output value. At the same time, it combines timing control (PWM modulation, time-division multiplexing) and closed-loop control (PID regulation) to ensure that the output current / voltage accurately matches the target parameters and adapts to the differentiated stimulation requirements of multi-region electrodes.

[0062] In practice, hBMSCs (P3 generation) were inoculated into a platinum-coated culture flask 1 containing a recessed culture region 2. Culture region 2 was divided into 9 independent regions (R) using a 90×90 pixel grid. The grid size was precisely converted to an actual physical region of 120μm×120μm using a pixel physical scale conversion factor k=1.2μm / pixel (assuming 90 pixels × 1.2μm / pixel = 108μm, but retaining the target physical size of 120μm×120μm as set by the user; accurate positioning was ensured through the grid division logic). The acquisition module was equipped with a 20x high numerical aperture (NA=0.75) infinity-corrected objective lens, a 480nm band 16-level brightness adjustable LED ring light source, and a 2048×2048 pixel back-illuminated sCMOS sensor (assuming a pixel size of 6.5μm). Data was transmitted at high speed via a 10GigE interface. The timing unit used a real-time clock chip with a timing accuracy of 0.5μs to ensure the synchronization accuracy between the timestamp and the acquisition position. The electrical stimulation module contains 4 platinum parallel electrodes and 8 sets of platinum DC adapter components arranged in a mesh (adapting to the gating control of 9 regions). A platinum metal sheet 3 is embedded on the side of culture region 2, and a fluorescent fiber optic temperature measuring device is attached to the outside, assuming a temperature monitoring accuracy of ±0.1℃ to ensure a stable cell culture environment (assumed to be 37℃±0.5℃). After cell seeding, the acquisition module acquires macro images of culture region 2 every 6 hours, synchronously linking the timestamp with the precise acquisition position of the 9 regions. In the image preprocessing stage, a 5×5 Gaussian filter kernel (assumed σ=1.0) is used for noise reduction, and 8×8 grid adaptive histogram equalization (assumed clipLimit=2.0) is used to enhance cell edge contrast, according to the formula... The original grayscale image is normalized to the 0-255 range to effectively correct for uneven illumination. In the segmentation stage, the Otsu algorithm is used to automatically calculate the optimal threshold (based on the maximum inter-class variance criterion), combined with the Canny algorithm (assuming a low threshold of 50 and a high threshold of 150) to complete the broken edges. For overlapping cells, seed points are located using distance transformation, and then the watershed algorithm is used for precise separation, avoiding feature extraction errors caused by cell contour adhesion. In the feature extraction stage, effective cell contours with an area of ​​50-5000 pixels and a roundness of 0.2-1.0 are retained according to the selection criteria. The length, width, and total area of ​​cells in each region are calculated. The cell nuclear area is obtained through a cell nuclear segmentation algorithm and compared with the total cell area to obtain the nucleocytoplasmic ratio (assuming that the key characteristic of osteogenic differentiation is approximately 0.3 in the proliferative phase and approximately 0.2 in the differentiation and maturation phase). The number and length of cell processes are counted (assuming that mature osteoblasts need to meet the requirement of ≥3 processes and an average length ≥10μm).

[0063] In permutation feature analysis, according to formula²² (M=90 pixels, which is the grid side length) Calculate cell density and classify it into three categories: low density (assumed to be <5 cells / mm²), medium density (assumed to be 5-20 cells / mm²), and high density (assumed to be >20 cells / mm²). By extracting the major axis of the smallest bounding rectangle of the cells, calculate the angle between it and the horizontal axis of the image, and statistically analyze the distribution ratio in 15° intervals. If the ratio is ≥60%, it is determined to be oriented, which provides a basis for subsequent stimulation parameter matching.

[0064] The analysis module for differentiation stage determination and parameter optimization is based on a preset initial mapping table. It matches the extracted morphological and arrangement features with the three stages of osteogenic differentiation (assumed to be P: proliferative phase p1, early differentiation phase p2, and differentiation maturation phase p3). The matching results are assumed to be as follows: in phase p1, cells are low-density, randomly arranged, and have ≤2 protrusions; in phase p2, cells are medium-density, locally oriented, and have 3-5 protrusions; and in phase p3, cells reach high density, globally oriented, and form a network structure.

[0065] according to The construction is performed, where the osteogenic differentiation-specific parameters are assumed to be: , , Stimulus intensity sensitivity coefficient , Quantitative calculations were performed based on the target morphology of osteoblasts (assuming a nucleocytoplasmic ratio of 0.2, ≥5 processes, and ≥80% directional arrangement). Constraints were strictly tailored to the osteogenic differentiation characteristics of hBMSCs, assuming: minimum effective stimulation intensity. V / cm, maximum endurance strength at each stage V / cm, V / cm, V / cm; Minimum threshold for morphological matching Stimulus type fit set (Pulse electricity only) , (Direct current / pulse current); Maximum permissible difference in stimulation intensity between adjacent regions V / cm.

[0066] When using a genetic algorithm for optimization, the population size is set to 50, and the initial optimization unit is set to... The encoding (9 sets of parameters corresponding to 9 regions) and the fitness function are: .

[0067] After 100 iterations (assuming a crossover probability of 0.8 and a mutation probability of 0.05), convergence is achieved after the rate of change of the optimal fitness value is less than 0.1% for 10 consecutive generations. The optimal parameters at each time point are assumed to be: 0h (p1 stage) uniformly adopted across the 9 regions. V / cm, (Pulsed electricity); 12h (P2 stage) high-density region (e.g., regions 3, 7) V / cm, medium-low density areas V / cm, ; 24h (P3 stage) oriented arrangement regions (such as regions 5 and 9) V / cm, (Direct current), other areas V / cm, .

[0068] The electrical stimulation module executes the differentiation regulation. After receiving stimulation commands from the analysis module, the main control unit analyzes the target parameters of the nine regions, converts the digital signals into analog reference values ​​via a DAC, and amplifies and outputs them through an H-bridge drive circuit (IR2104 gate driver, IRF540N power MOSFET). During the proliferation phase (0-12h), pulsed electrical stimulation is output to all regions through parallel electrode 4. The DC adapter adapts to the load impedance of each region in real time to compensate for parameter attenuation caused by changes in the culture medium impedance. During the differentiation and maturation phase (assumed to be 24-72h), the corresponding DC adapter is selected according to the region to accurately output DC electrical stimulation to the oriented regions. The sampling feedback unit collects the actual output parameters in real time through a Hall current sensor and a voltage divider sensor. After ADC conversion, the parameters are sent back to the main control unit. The deviation is corrected by a PID algorithm to ensure that the current / voltage error is controlled within ±1%. During the process, the fluorescent fiber optic thermometer collects temperature data every 30 seconds. The data is then transmitted to the acquisition module, converted into temperature data by the AD converter, and synchronized to the analysis module. When the temperature in a certain area exceeds 37.5℃, the analysis module immediately issues an adjustment command, and the electrical stimulation module automatically reduces the stimulation intensity in that area by 0.5V / cm until the temperature returns to the normal range, thus ensuring cell activity.

[0069] In Example 2, the analysis module stores a temperature mapping table containing differentiation stages, differentiation directions, and culture temperature ranges. Based on the current differentiation stage and the differentiation direction corresponding to the preset target morphology, a culture temperature range is matched in the temperature mapping table as a reference temperature range. This reference temperature range is then combined with the current cell position distribution at each differentiation stage to obtain the reference temperature distribution, which is then sent to the electrical stimulation module. The parameter matching mapping table also stores heat generation data for each electromagnetic device under different operating parameters. The electrical stimulation module calculates the total heat generation data in each culture region 2 based on the current operating status of each electromagnetic component. The total heat generation data is compared with the temperature distribution data to obtain actual heat dissipation data. The reference temperature distribution and the temperature distribution data are compared to obtain the temperature adjustment range. The temperature adjustment range, heat generation data, and actual heat dissipation data are combined to adjust the operating parameters of each electrical stimulation component without changing the stimulation range or intensity. The parameter matching mapping table stores multi-dimensional associated data categorized by the unique identification number of the electromagnetic device. It includes the matching relationships between stimulus type (S, assuming 1 represents direct current and 2 represents pulsed current, a discrete code), stimulus intensity (E, unit V / cm), and operating current (I, unit mA) and operating voltage (U, unit V), satisfying the formula... (Where d is the electrode spacing, a fixed physical parameter to ensure that the stimulation intensity is rigidly bound to the operating parameters).

[0070] It also includes the basic heat generation power of each electromagnetic device under different combinations of operating parameters (I, U, S, duty cycle D). (Unit: W), D ranges from 0 to 1 (during continuous operation). ), Calibration was performed under standard environmental conditions (assuming 37°C, 50% humidity, and standard culture medium). Auxiliary correction parameters were also included, such as energy conversion efficiency (…). Dimensionless, determined by the material and structure of the device, such as platinum. ) and heat loss coefficient ( (Unit: W / ℃, characterizing the heat loss characteristics of the device itself).

[0071] Let the total number of culture regions 2 be R (region number). The total number of electromagnetic devices is M (device number). ); The set of electromagnetic devices covering the r-th cultivation region 2, The ratio of the area covered by the stimulation of the m-th device to the area of ​​the r-th region (when a single device covers a single region). First, calculate the actual heat output of a single unit: ,in The actual heat output power (in W) of device m. This represents the base heat production capacity of device m under the current operating parameters. , , , These are the real-time operating current, voltage, stimulation type, and duty cycle of device m, respectively. Let m be the energy conversion efficiency of device m. Let m be the heat loss coefficient of device m. The real-time ambient temperature is expressed in °C. The target temperature is preset (in °C). Then, the total heat output of the superimposed region is calculated: ,in Let be the total heat production power (in W) of the r-th region. This indicates that the heat generated by all devices covering the area is accumulated.

[0072] Derivation of actual heat dissipation power based on the principle of thermal balance (Unit: W): First, collect the real-time temperature of the r-th region at time t. (Unit: °C) Temperature at the previous sampling time (Unit: °C) For the sampling interval (e.g., 30s), according to the formula Calculate the rate of temperature change (Unit: °C / h); then retrieve the pre-calibrated regional working fluid comprehensive heat capacity. (Unit: Wh / ℃, working medium includes culture medium, culture carrier, and cells) and environmental heat dissipation correction factor (Unit: W / ℃, representing the effect of temperature and humidity on heat dissipation), according to the formula Calculation, where The total heat production capacity of the region, The heat consumed to heat the working fluid This accounts for heat loss due to environmental heat. A preset reference temperature distribution is used, including the target temperature. (unit: °C) and allowable fluctuation threshold (Unit: °C, e.g., °C) ℃ According to the formula Calculate temperature deviation (Unit: °C) When the temperature is too high, adjust the range. (Requires at least cooling) ); When the temperature is too low, adjust the range. (At least heating is required) ); No adjustment is needed. The stimulation range (device gate state remains unchanged) and stimulation intensity (U fixed, E unchanged) are not altered; only timing parameters (D or on / off cycle) are adjusted. The target actual heat output of device m is determined based on the heat balance objective. (Unit: W), based on The positive correlation with D, according to the formula Calculate the adjusted duty cycle ,in The duty cycle before adjustment. This represents the actual heat output before adjustment. After adjustment, temperature will be continuously monitored, and adjustments will be made in increments of 0.01. until the temperature stabilizes at Within the range.

[0073] In practice, the targeted differentiation of stem cells is the goal, and precise temperature control is achieved through the coordinated regulation of the analysis module and the electrical stimulation module. The analysis module stores a pre-defined temperature mapping table, which associates differentiation stage, differentiation direction, and corresponding culture temperature range. During actual operation, the analysis module first identifies the current differentiation stage of the cells, and, combined with the differentiation direction corresponding to the preset target morphology, matches the appropriate culture temperature range from the mapping table as a reference temperature range. Then, based on the regional distribution characteristics of cells at each differentiation stage, the reference temperature range is allocated to the corresponding culture region 2, forming a complete reference temperature distribution, which is simultaneously sent to the electrical stimulation module.

[0074] The parameter matching mapping table of the electrical stimulation module is categorized by the unique identification number of the electromagnetic device. It stores the matching relationship between stimulation type, stimulation intensity, and operating current and voltage, as well as the basic heat generation data under different combinations of operating parameters for each device. It also includes auxiliary correction parameters such as energy conversion efficiency and heat loss coefficient. When calculating the total heat generation, the actual heat generation of each individual device is first calculated based on the real-time operating status of each electromagnetic device and the correction parameters. Then, the heat generation of all devices covering the region is summarized to obtain the total heat generation data for each region.

[0075] Based on the principle of thermal balance, the electrostimulation module acquires real-time temperature and changes in each region through a temperature monitoring device. Combined with pre-calibrated comprehensive heat capacity of the working fluid and environmental heat dissipation correction coefficients, it derives the actual heat dissipation data for each region. By comparing the reference temperature distribution with the real-time temperature distribution, it determines whether the temperature is too high, too low, or within acceptable limits, thus clarifying the direction and requirements for temperature adjustment. While maintaining the stimulation range and intensity unchanged, the electrostimulation module only adjusts timing parameters such as the duty cycle or on / off cycle of the electromagnetic devices. Based on the thermal balance target, the target heat production of each device is determined, and the adjusted parameter values ​​are calculated by considering the correlation between current heat production and timing parameters. After adjustment, the temperature is continuously monitored, and small adjustments are made to stabilize the temperature in each region within the reference temperature range, ensuring the temperature stability of the cell differentiation environment.

[0076] Example 3

[0077] It also includes a culture medium supply circuit, which is electrically connected to the analysis module. The cell morphology evolution model also includes nutrient consumption rate data (preset by the technician). When the analysis module matches the stimulus intensity, it acquires the nutrient consumption rate data. When the difference between the reference temperature distribution and the temperature distribution data is greater than the preset temperature adjustment threshold (set by the technician according to the cell culture temperature range), the area greater than the preset temperature adjustment threshold is taken as the temperature adjustment area. A temperature adjustment command is generated based on the temperature adjustment area and sent to the culture medium supply circuit and the acquisition module. The culture medium circuit lowers or increases the temperature of the culture medium supplied to the higher temperature area and adjusts the culture medium supply volume. After receiving the temperature adjustment command, the acquisition module increases the acquisition frequency of macro images within the distance change range preset by the technician, according to the increase rate preset by the technician.

[0078] The culture medium loop, acting as an auxiliary execution unit for temperature regulation, forms a collaborative temperature control mechanism with the timing parameter adjustment of the electrical stimulation module, implementing regulation for areas with high temperatures. When the electrical stimulation module or analysis module determines that the temperature of a certain culture region 2 exceeds the upper limit of the reference temperature range, it immediately sends a temperature regulation command to the culture medium loop. Upon receiving the command, the loop control prioritizes reducing the temperature of the culture medium supplied to that high-temperature region through the built-in temperature control unit (such as a micro-cooling component). The adjustment range is positively correlated with the temperature deviation; for example, when the temperature exceeds the threshold by 0.5-1℃, the culture medium temperature decreases by 0.5-1℃; when the threshold exceeds by >1℃, the temperature reduction is increased to 1-1.5℃ to ensure rapid suppression of the temperature rise trend. Simultaneously, the culture medium supply volume is dynamically adjusted. For high-temperature regions, the supply is increased according to a preset proportion of the current supply volume (technicians typically set it between 10%-30%, depending on the cell differentiation type), enhancing heat dissipation efficiency by increasing the culture medium flow rate. If multiple regions are simultaneously high-temperature, the supply volume is allocated according to the regional temperature deviation weight, with regions with larger temperature deviations receiving more culture medium supply to avoid localized heat accumulation. During the adjustment process, the outlet temperature of the culture medium and the temperature of the cells in the region are collected in real time to form a closed-loop feedback, dynamically correct the cooling range and supply volume, and prevent temperature fluctuations caused by over-adjustment.

[0079] When multiple culture regions 2 simultaneously experience temperatures exceeding the upper limit of the reference range, all culture regions 2 that meet the condition of real-time temperature > upper limit of the reference temperature are selected to form an over-temperature region set. ( , (Region number). Real-time temperature of the region. Real-time data is collected by a fluorescent fiber optic temperature measuring device (an auxiliary component of the acquisition module), with a sampling interval of 30 seconds, directly obtaining the time of each area. Actual temperature data; regional reference temperature upper limit The analysis module matches the data from the temperature mapping table based on the cell differentiation stage and direction (referencing the temperature range). (the upper limit of the value), for example, the osteogenic differentiation and proliferation phase of stem cells at ℃. .when At that time, the area was determined to be an overheated area.

[0080] For each overheated zone Calculate the absolute deviation value exceeding the allowable upper limit using the formula: , The steps are the same as above, with data collected in real time from the temperature sensor; Similar to the steps above, the temperature mapping table is matched from the analysis module; Unitless, only quantifies the excess temperature increase, for example, ℃. ℃ Then ℃ .

[0081] Collection of overheated regions All regions Summing, we get the total deviation: , Unitless; reflects the total temperature rise load across all overheated zones, e.g., 3 overheated zones. If the deviations are 0.4℃, 0.6℃, and 0.5℃ respectively, the total deviation is 1.5℃.

[0082] A single overheated area Total Deviation The ratio of these values ​​is used as the weight of that region. , The value ranges from 0 to 1. The sum of the weights of all over-temperature areas is 1, and the larger the deviation, the higher the weight. For example, if the total deviation is 1.5℃, the weight of a certain area is ℃. ,but (40% weight).

[0083] Based on the sum of the basic supply volumes of each overheated zone, the total increment is calculated proportionally: , : The sum of the basic supply volumes for all overheated zones (unit: mL / h). Preset the basic supply for a single region (set before the experiment according to the area and cell density of culture region 2, for example, a 120μm×120μm region). ), ( ); Incremental ratio coefficient (value ranges from 10% to 30%), determined by the degree of overheating: Total deviation in °C hour At 1-2℃ When >2℃ Additional supply volume in a single overheated zone: Final supply volume of a single overheated zone: ;like Then take (To avoid excessively rapid flow that could affect cell adhesion). The additional supply obtained in a single region, for example , ,but ; The final supply volume for a region is determined to ensure that regions with large deviations receive more resources. For example, a region with a basic supply of 10 mL / h will ultimately receive 12.4 mL / h. The above steps are repeated every 30 seconds to recalculate. , and According to the temperature change rate of each region ( Adjust the incremental ratio until all areas To restore the basic supply volume. Temperature change rate (°C / s): Calculated from the difference between two consecutive temperature data acquisitions, used to judge the temperature control effect. For example, < This indicates that the temperature has dropped and can be maintained. ; This indicates insufficient temperature control and needs to be improved. .

[0084] The culture medium circuit has local nozzle components above the parallel passages of each electrode. These nozzle components are fixedly connected to the inner wall of the culture flask 1 at the upper end of the culture area 2. The culture medium circuit includes separate passages between each local nozzle component and the individual passages. Each individual passage has a control valve that controls the volume of culture medium delivered to the local nozzle component, as well as a temperature regulating device that is connected to each individual passage. Each temperature regulating device and control valve is electrically connected to the analysis module. The analysis module controls the temperature of the culture medium sprayed to each position in the culture area 2 by using the unique identification number of the temperature regulating device and the corresponding position of the local nozzle component. It controls the volume of culture medium sprayed to each position in the culture area 2 by using the control valve number and the corresponding control valve position (a flow rate detection sensor is set up, and the culture medium volume is calculated by combining the collected flow rate and the opening time of the control valve).

[0085] The analysis module counts the number of temperature adjustments and the corresponding adjustment time based on the time the temperature regulation device is connected to the current. The module designates the region that meets the preset target morphology at the current differentiation stage as successful culture region 2, and the time corresponding to meeting the preset target morphology at the current differentiation stage as the successful culture time. It acquires the operating parameters of each electrical stimulation component of the electrical stimulation module within the successful culture time, constructs an electrical stimulation dynamic model of the electrical stimulation type and intensity distribution at each time within each culture region 2, acquires temperature distribution data within the successful culture time, constructs a temperature dynamic model, acquires the culture medium supply time and supply location within the successful culture time, and constructs a supply dynamic model. It also acquires the morphological and arrangement characteristics of cells in macro images within the successful culture time and constructs a morphological dynamic model. Finally, it performs a collaborative analysis of the parameter change trends and correlations of the electrical stimulation dynamic model, temperature dynamic model, supply dynamic model, and morphological dynamic model by aligning them over time.

[0086] Among them, the dynamic model of electrical stimulation quantifies the distribution of stimulation parameters in culture region 2 over time, expressed as: , (Total number of cultivation areas 2) (Total training time successfully); (V / cm) represents the region time The intensity of the stimulus, to satisfy ( (The differentiation stage) (1 = DC, 2 = pulse) represents the stimulation type, satisfying... Discrete data interpolation completion: ( ), The k-th discrete time node (e.g., 0h, 6h, 12h, etc.) is a known time reference point; For the first Each discrete time point is The next time reference point. The temperature dynamics model reconstructs the spatiotemporal evolution of temperature, constructed based on thermal equilibrium: , (°C) represents the region. time temperature, The initial temperature;

[0087] (Total heat production power) The unique number for culture region 2 (1, 2, ..., R, where R is the total number of culture regions 2); For any time (a variable in the time dimension, used to represent a point in time during the cultivation process); This indicates that the electromagnetic devices covering the r-th culture region 2 are cumulatively calculated; m is the unique number of the electrical stimulation component (e.g., 1, 2, ..., M, where M is the total number of electrical stimulation components). The set of electromagnetic devices covering the r-th culture region 2 (i.e., all devices that can have a thermal effect on region r). This represents the ratio of the effective range of the m-th electrical stimulation component to the area of ​​the r-th culture region 2. The m-th electromagnetic device is at time... The actual heat production power (effective heat production after deducting the heat loss of the device itself). (Actual heat dissipation power); The point above T represents the derivative of temperature with respect to time, indicating the temperature of the r-th region at time T. The rate of temperature change; The heat dissipation coefficient is set by technicians based on the heat conduction characteristics of the culture device and the heat exchange efficiency of the environment, reflecting the heat dissipation capacity of the area to the outside world. The ambient temperature is the reference temperature of the external environment in which the culture device is located. (Wh / ℃) represents the total heat capacity of the working fluid, and the data comes from the measured temperature distribution data.

[0088] Supply dynamics model quantifies the regulation of culture medium supply: , (mL / h) is the basic supply volume. This is the new addition; ,in, ( (The increment coefficient is 10%-30%), where r is the unique identifier for culture region 2. Let t be the total basic supply volume of all overheated regions (the total normal supply before any increase). (Temperature deviation weight). Let be the absolute temperature deviation of the r-th region at time t (the portion of the actual temperature that exceeds the upper limit of the reference temperature). This represents the summation of the absolute temperature deviations of all overheated regions at time t. This is the set of overheated regions at time t (all culture regions 2 where the actual temperature exceeds the upper limit of the reference temperature).

[0089] Morphological dynamic models quantify the temporal evolution of cell features: , : Represents the first In cultivation region 2 at time The morphological dynamic model is a comprehensive representation of the cellular morphological characteristics of this region; : The unique number of cultivation area 2 (e.g., 1, 2, 3...); A point in the cultivation process; : No. Each region at time The cell feature vector contains four core morphological / arrangement features; , , , These correspond to four core parameters in sequence: nucleocytoplasmic ratio, protrusion characteristics, arrangement direction characteristics, and cell density characteristics. (Nucleo-cytoplasmic ratio) : No. Each region at time The nucleus-cytoplasm ratio (a key indicator reflecting cell growth / differentiation status); : No. Each region at time The area of ​​the nucleus of a single cell (obtained by feature extraction from macro images); : No. Each region at time The total area of ​​a single cell (including cytoplasm and nucleus, obtained by feature extraction from macro images). (Protrusion feature) : No. Each region at time Cellular protrusion characteristics (comprehensively reflecting the degree of cell differentiation, such as the protrusion development during osteogenic differentiation of stem cells); : No. Each region at time The number of protrusions in a single cell (obtained by feature extraction from macro images); : No. Each region at time The average length of the protrusions in a single cell (obtained by feature extraction from macro images); the square root sign serves to comprehensively quantify the number of protrusions and their average length, avoiding the influence of a single parameter's deviation on feature accuracy. (Arrangement direction) : No. Each region at time The characteristics of cell arrangement direction (reflecting the directional growth state of cells); After dividing the cell arrangement angle into 12 intervals of 15° (set by technicians based on the angular distribution characteristics of cell arrangement and the practicality of quantitative analysis, this is a hypothetical value), the proportion of the number of cells in each interval to the total number of cells in that area (obtained by the orientation detection function of macro images). The maximum value among these proportions represents the main orientation of cells in that region. (density), : No. Each region at time Cell density (reflecting the density of cells in a region); : No. Each region at time The total number of cells (obtained by the counting function of the macro image); Pixel-to-physical scale conversion factor (used to convert the pixel units of an image to actual physical units, such as...) (pixels) : No. The actual physical area of ​​each culture region 2 (e.g., 120μm × 120μm). : A coefficient that converts pixel area to physical area, used to represent the uniform number of cells within a region and the unit dimension of the region's area.

[0090] When aligning models in time, the incubation start time is used as the starting point. The time unit is uniformly set to hours; discrete data is interpolated and completed using a minimum sampling interval of 30 seconds (set by technicians according to the dynamic rate of change in cell culture, this is an assumed value); and a uniform timestamp is bound to each model parameter. ,form Data quadruples under coordinates. When performing model co-analysis (correlation quantification formula), analyze trend correlation: , , For the parameter sequence to be analyzed (e.g.) and ), The correlation was determined to be strong. (Calculation results) It is the Pearson correlation coefficient, and its value ranges from... between: The closer the value is to 1, the stronger the linear trend correlation between the two parameter sequences. A positive value indicates a positive correlation between the two sequences (when one parameter increases, the other also increases synchronously); a negative value indicates a negative correlation (when one parameter increases, the other decreases synchronously). In the formula, X and Y are two parameter sequences to be analyzed (e.g., a time-varying sequence of electrical stimulation intensity, a time-varying sequence of temperature in culture region 2, or a time-varying sequence of cell morphology characteristics); t is the number of the time sampling point, and n is the total number of sampling points for these two parameter sequences. The numerator of the formula... This reflects the degree of covariance between X and Y: where It is the value of the parameter sequence X at the t-th sampling point. It is the average value of all sampled points in the X sequence; It is the value of the parameter sequence Y at the t-th sampling point. This is the average value of the Y series. If X and Y frequently deviate from their respective averages simultaneously, this value will increase, meaning that their trends are more consistent. The denominator of the formula... , which is the product of the standard deviations of the X and Y sequences, serves to standardize the degree of covariance of the molecules, thus constraining the final correlation coefficient to a certain range. Within a given range, it is easy to intuitively judge the strength of the correlation.

[0091] Analyze timing consistency: This formula quantifies the temporal synchronicity of changes in two parameters, primarily reflecting whether the response rhythm of the control operation and its effect matches. Among these, This is the quantization result for temporal consistency; n is the total number of sampling points for the two parameter sequences; t is the sampling point number; The rate of change of the representative parameter X at the t-th sampling point over time (e.g., the rate of adjustment of the intensity of electrical stimulation). The value represents the rate of change of parameter Y at the t-th sampling point (e.g., the rate of temperature change in culture region 2). The formula first calculates the absolute value of the difference between the rates of change of the two parameters, and then averages the results over all sampling points. The smaller the value, the closer the changing rhythms of the two parameters are (e.g., the temperature changes synchronously after the electrical stimulation intensity is adjusted), which reflects a higher response efficiency and more precise effect of the control strategy. The smaller the value, the more consistent the timing of parameter changes, reflecting the effectiveness of the control.

[0092] Analysis of correlation contribution : This formula is used to quantify the influence weight of parameter X on parameter Y, clarifying the degree of influence of different regulatory factors on the target indicator. Among them, It represents the correlation contribution of parameter X to parameter Y; It is the total variance of parameter Y (reflecting the degree of fluctuation of Y itself); It is the conditional variance of parameter Y after controlling parameter X to remain stable (reflecting the residual volatility of Y after excluding the influence of X). The result of this formula takes values ​​in... Between these values, the larger the value, the greater the influence of X on Y (for example, the higher the contribution of electrical stimulation intensity to the temperature of culture region 2, meaning that adjusting electrical stimulation is the core means of temperature control). This result can provide a basis for subsequent optimization of control strategies, that is, prioritizing the adjustment of parameters with high contribution can improve control efficiency.

[0093] Based on the collaborative analysis results, the analysis module generates parameter write-back adjustment instructions, writes back and optimizes the electrical stimulation type, electrical stimulation time, electrical stimulation intensity, culture temperature range, and nutrient consumption rate data, and associates the optimized parameters with the location information of the successful culture area 2 in culture area 2, storing it as a default template in the built-in database; when a new batch of culture is started, the analysis module generates a template call instruction, and directly applies the default template to the corresponding culture area 2 of culture bottle 1.

[0094] When applying the default template in a new batch of culture, the analysis module collects electrical stimulation parameters, temperature distribution data, culture medium supply data, and cell morphology and arrangement characteristics data in real time during the actual culture process. It compares these data with the parameters of each dynamic model corresponding to the default template in real time and calculates the parameter deviation. When the deviation exceeds a preset threshold, a real-time fine-tuning command is generated to dynamically adjust the electrical stimulation type, culture temperature, and culture medium supply volume in the corresponding region. After the new batch of culture is completed, the analysis module merges the complete data of the successfully cultured region 2 with the original default template data, performs a new multi-model collaborative analysis, generates a template update command, iteratively optimizes the parameters and regional relationships of the default template, and stores the results. During the actual culture process, the analysis module simultaneously collects in real time the electrical stimulation parameters (including the time distribution of stimulation type and intensity), regional temperature distribution data fed back by the fluorescence fiber optic thermometer, the supply volume of the culture medium circuit, time-series data, and cell morphology (nucleocytoplasmic ratio, protrusion characteristics) and arrangement characteristics (orientation, density) data output by the acquisition module for each culture region 2. For each culture region 2 and each time point, the analysis module compares the real-time collected parameters with the baseline parameters of the corresponding dynamic model in the default template point by point, and calculates the parameter deviation value: the deviation value is the absolute value of the difference between the actual parameter and the template baseline value (for example, electrical stimulation intensity deviation = |actual intensity - template baseline intensity|, temperature deviation = |actual temperature - template baseline temperature|).

[0095] The analysis module has built-in preset deviation thresholds (based on the stable range of historical culture, such as an electrical stimulation intensity deviation threshold of 0.2V / cm, a temperature deviation threshold of 0.3℃, and a cell morphology characteristic deviation threshold of 0.1V / cm). When the deviation value of any parameter in any region exceeds the corresponding threshold, the analysis module dynamically adjusts the electrical stimulation type (e.g., switching from pulsed electricity to direct current to change the heat production rhythm) and culture temperature (by adjusting the electrical stimulation intensity to change heat production, or by adjusting the supply temperature in conjunction with the culture medium circuit) for the deviation region. Simultaneously, it adapts the increase in culture medium supply volume according to the region's deviation weight. If the temperature exceeds 0.5℃, pulsed electricity is switched to direct current (for stable heat production); if the temperature exceeds 0.3℃, direct current is switched to pulsed electricity (to increase instantaneous heat production). The electrical stimulation intensity is adjusted according to the deviation: ±0.1V / cm for 0.3-0.5℃, ±0.2V / cm for 0.5℃ above the deviation; simultaneously, the culture medium circuit is linked: the supply medium temperature is lowered by 0.5-1℃ for over-temperature and raised by 0.3-0.8℃ for under-temperature. Regions with parameter deviations exceeding the threshold are selected. The absolute deviation of a single region (the absolute value of the difference between the actual value and the default template baseline value) is calculated. The deviation of a single region is divided by the total deviation of all deviation regions to obtain the weight (values ​​range from 0 to 1, with a weight sum of 1). Based on the basic supply volume of the deviation regions, the total increment is calculated at 10%-30% (the larger the deviation, the higher the percentage). The increase in supply per region = total increment × corresponding weight; the final supply is constrained to not exceed 1.5 times the basic volume to avoid excessive flow rate affecting cell adhesion. After the fine-tuning command is issued, the analysis module continuously collects the adjusted parameter data and repeats the comparison process with the default template until the parameter deviation falls back to the threshold range. The electrical stimulation module adds an impedance acquisition circuit. While constructing the electrical stimulation dynamic model, the analysis module generates electrode function switching commands to control the dot-matrix distributed metal sheet electrodes (such as...). Figure 4 (As shown) It alternately switches between the current injection electrode and the voltage detection electrode.

[0096] The impedance acquisition circuit acquires multiple sets of impedance signals during electrode switching and transmits them to the analysis module. The analysis module calls the electrical impedance tomography (EIT) algorithm to invert the conductivity distribution of the culture medium and extract data from cell aggregation regions. This cell aggregation region data is then integrated into a morphological dynamic model and subjected to four-dimensional collaborative analysis with the electrical stimulation dynamic model, temperature dynamic model, and supply dynamic model. During parameter write-back optimization, the electrical stimulation intensity gradient and culture medium supply density are adjusted specifically for the corresponding regions based on the cell aggregation region data. The analysis module automatically triggers the electrical impedance tomography (EIT) algorithm, first inputting a preset baseline conductivity value for the culture medium. , The conductivity of the pure culture medium (unit: Siemens / m, S / m) measured in the early stage of culture (before cells proliferate and aggregate in large quantities) is used as a benchmark for conductivity comparison, for example, a value of 1.2 S / m. Real-time voltage response data from the built-in electrode array of the culture device are acquired simultaneously. (Unit: Volt, V), where The cultivation area 2 is uniquely numbered (e.g., 1, 2...). , (Total number of regions) For specific moments in the cultivation process, a modified Newton-Raphson inversion algorithm is used, based on the mapping relationship between voltage response and conductivity, to calculate the values ​​for each cultivation region. At any moment Real-time conductivity (Unit: S / m) This ultimately generates a full-area conductivity distribution matrix containing conductivity data for all regions. A conductivity deviation judgment threshold is set: (Because cell aggregation leads to a significant increase in regional conductivity, this threshold was calibrated based on the difference in conductivity between pure culture medium and cell aggregation medium.)

[0097] Filter all regions that meet the above threshold conditions and mark them as a set of cell aggregation regions. ( (No unit, only represents a set of cluster area numbers); core data for each cluster area is recorded synchronously, including: area coordinates (physical location information, such as...) (Plane coordinates); conductivity deviation value : (Unit: S / m, quantifying the increase in conductivity caused by aggregation); initial cell density data of the cell aggregation region (preset cell seeding density at culture start, unit: cells / mm²), these three are correlated to form a complete cell aggregation region dataset. The conductivity deviation value of the cell aggregation region is then used to... As a new feature, the feature vector of the original dynamic model is updated, and the updated expression is: Cultivation area number 2 : Cultivation time; Nucleoplasm-to-mass ratio characteristics (unitless) , The area of ​​the cell nucleus. (total cell area); : Protrusion characteristics (no unit, , The number of protrusions. (meaning the average length of the protrusion). Arrangement direction characteristics (unitless, percentage of cells in the maximum angle range); Cell density characteristics (unit: cells / mm²) , For cell number, This is the pixel-to-physical scale conversion factor. (area of ​​the region) Cell aggregation characteristics (unit: S / m, directly assigned value) (Quantization of aggregation degree). Using a uniform sampling interval of 30 seconds, the updated morphological dynamic model... With dynamic model of electrical stimulation Temperature dynamic model Supply dynamics model The parameters are bound to the same timestamp. This ensures a one-to-one correspondence of four-dimensional parameters within the same timeframe and region. Clustering features are then calculated. With electrical stimulation intensity Pearson correlation coefficient (unit: V / cm) (determining the degree of linear association between cell aggregation and electrical stimulation), and With regional temperature (Unit: °C) temporal consistency (determining the synchronous response rhythm of aggregation and temperature control), and finally outputting a quantization matrix containing the correlation strength of four-dimensional parameters.

[0098] When the cell aggregation region meets the (High-aggregation state): The electrical stimulation intensity gradient between this region and adjacent regions is set to 0.05V / cm. This intensity gradient refers to the difference in electrical stimulation intensity between adjacent regions. Reducing the gradient can prevent excessive local electric field concentration leading to excessive heat generation and reduce abnormal temperature increases in the aggregation region. When the region meets... (Low-aggregation / non-aggregation state): The intensity gradient is set to 0.1V / cm to amplify the intensity difference between adjacent regions. The electric field guides cells to diffuse towards the low-aggregation area, thus balancing cell distribution in the region. Cell aggregation area ( ): Adjust the nutrient supply density (defined as the volume of culture medium supplied per unit area, unit: mL / mm²) to 1.2 times the baseline nutrient supply density. For example, if the baseline nutrient supply density is 0.5 mL / mm², adjust it to 0.6 mL / mm². By increasing the nutrient supply per unit area, the high nutrient demand of the aggregation area is matched, avoiding local nutrient deficiency. Low aggregation area ( To maintain a constant baseline supply density, avoid nutrient overload leading to non-specific cell aggregation and ensure normal cell growth and differentiation. After adjustment, repeat the above EIT inversion, aggregation region extraction, and four-dimensional synergistic analysis process every 30 minutes, iteratively optimizing the electrical stimulation intensity gradient and culture medium supply density parameters until the cell aggregation distribution meets the uniformity requirements of the target differentiation.

[0099] In practice, the analysis module is electrically connected to the culture medium supply circuit, and the nutrient consumption rate data preset by the technician is incorporated into the cell morphology evolution model. When matching and determining the electrical stimulation intensity, the analysis module simultaneously retrieves this nutrient consumption rate data. When there is a difference between the reference temperature distribution and the actual acquired temperature distribution data, and this difference exceeds the temperature adjustment threshold set by the technician based on the cell culture temperature range, the area is identified as a temperature adjustment zone. The analysis module then generates a temperature adjustment command and sends it to both the culture medium supply circuit and the acquisition module. Upon receiving the command, the culture medium circuit adjusts the temperature of the delivered culture medium through its built-in temperature control unit for areas with higher temperatures, while dynamically changing the culture medium supply volume. Upon receiving the command, the acquisition module increases the acquisition frequency of macro images within a preset distance range according to the increment preset by the technician.

[0100] The culture medium circuit, acting as an auxiliary execution unit for temperature regulation, forms a collaborative temperature control mechanism with the timing parameter adjustment of the electrical stimulation module. When the temperature of a certain culture region 2 exceeds the upper limit of the reference temperature range, the culture medium circuit prioritizes reducing the temperature of the culture medium supplied to that region through temperature control units such as micro-cooling components. The cooling rate varies accordingly with the temperature deviation to quickly suppress the temperature rise. Simultaneously, the supply volume is increased according to a preset ratio, enhancing heat dissipation efficiency by increasing the culture medium flow rate. If multiple regions are simultaneously at high temperatures, the supply volume is allocated according to the temperature deviation weight of each region, with regions having larger temperature deviations receiving more culture medium to avoid localized heat accumulation. During the adjustment process, the culture medium outlet temperature and the regional cell temperature are collected in real time to form a closed-loop feedback, dynamically correcting the cooling rate and supply volume to prevent over-adjustment and temperature fluctuations. The analysis module defines regions that meet the preset target morphology at the current differentiation stage as successfully cultured region 2, with the corresponding time being the successful culture time. Based on the relevant data within the successful culture time, dynamic models for electrical stimulation, temperature, supply, and morphology are constructed respectively. The electrical stimulation dynamic model quantifies the distribution of electrical stimulation type and intensity at different times in each culture region; the temperature dynamic model, based on the principle of thermal equilibrium, reconstructs the spatiotemporal evolution of temperature; the supply dynamic model records the supply time and location of the culture medium; and the morphological dynamic model extracts the morphological and arrangement characteristics of cells.

[0101] Subsequently, the four models were time-aligned to ensure that parameters in the same time period and region corresponded one-to-one. Then, trend correlation analysis, temporal consistency analysis, and correlation contribution analysis were used to uncover the changing trends and intrinsic relationships of the parameters in each model. Trend correlation analysis quantifies the linear correlation between two parameter sequences, temporal consistency analysis determines the temporal synchronicity of parameter changes, and correlation contribution analysis clarifies the influence weights between different parameters. Based on the collaborative analysis results, the analysis module generates parameter write-back adjustment instructions to optimize the electrical stimulation type, time, intensity, culture temperature range, and nutrient consumption rate data. The optimized parameters are then associated with the location information of the successfully cultured region 2 and stored as a default template in the built-in database. When a new batch of culture is started, the analysis module calls this default template and applies it to the corresponding culture region 2. During the new batch of culture, the analysis module collects electrical stimulation parameters, temperature distribution data, culture medium supply data, and cell morphology and arrangement characteristics data in real time, comparing them point-by-point with the model parameters in the default template to calculate parameter deviation values. When the deviation value exceeds a preset threshold, a real-time fine-tuning instruction is generated to dynamically adjust the electrical stimulation type, culture temperature, and culture medium supply volume in the corresponding region. During adjustment, the type of electrical stimulation is switched according to the temperature deviation of the region, and the intensity of electrical stimulation is adjusted according to the degree of deviation. At the same time, the temperature of the supply medium is adjusted in conjunction with the culture medium circuit. The regional deviation weight is obtained by screening the deviation region, calculating the absolute deviation of a single region and normalizing it. The increase in culture medium is allocated according to this weight, and the final supply volume does not exceed the upper limit of the baseline value to avoid affecting cell adhesion. After fine-tuning, data is continuously collected and compared repeatedly until the deviation falls back to within the threshold.

[0102] The electrical stimulation module incorporates an impedance acquisition circuit. When constructing the dynamic model of electrical stimulation, the analysis module generates electrode function switching commands, controlling the lattice-distributed metal sheet electrodes to alternately serve as current injection electrodes and voltage detection electrodes. The impedance acquisition circuit acquires multiple sets of impedance signals and transmits them to the analysis module. The analysis module uses an electrical impedance tomography algorithm to invert the conductivity distribution of the culture medium, filters out cell aggregation regions, and records the coordinates, conductivity deviation values, and initial cell density data of these regions, forming a cell aggregation region dataset. The conductivity deviation values ​​are incorporated as a new feature into the morphological dynamic model, and the updated morphological model undergoes four-dimensional collaborative analysis with the other three major models. During parameter write-back optimization, the electrical stimulation intensity gradient and culture medium supply density of the corresponding regions are adjusted based on the cell aggregation data: the electrical stimulation intensity gradient between high-aggregation regions and adjacent regions is reduced, and the culture medium supply density is increased; the intensity gradient in low-aggregation regions is expanded to maintain the basic supply density and guide balanced cell distribution. After adjustment, the above process is repeated at intervals to iteratively optimize the parameters until the cell aggregation distribution meets the uniformity requirements of the target differentiation. After the new batch of culture is completed, the analysis module will merge the complete data of the actual successfully cultured region 2 with the original default template, re-perform multi-model collaborative analysis, iteratively optimize the parameters and regional correlations of the default template and store them to achieve continuous improvement of the template.

[0103] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A direct current stimulation stem cell culture device, characterized in that, include: The acquisition module acquires macro image information within a preset culture range, associates the acquisition time with the macro image information as a timestamp, extracts cell contour features from the macro image information, and identifies and marks the arrangement features from the cell contour features; The analysis module acquires the acquisition location and acquisition parameters of macro image information, processes them to obtain the size information of the macro image information, and quantifies the size information of the cells by combining the relative positional relationship between cell contour features and macro image information to obtain morphological features; it stores an initial mapping table between cell differentiation stages and morphological and arrangement features, and matches the morphological and arrangement features of cells in macro image information with the initial mapping table to obtain the positional distribution of cells at each differentiation stage; An electrical stimulation module includes an electrical stimulation component, which is used to provide direct current stimulation or pulse stimulation to a preset culture range; The analysis module stores cell morphology evolution models of cells with different morphologies at different differentiation stages under different stimulation intensities, stimulation types, and stimulation times. The analysis module obtains the current differentiation stage of the cells, combines the current differentiation stage with the preset target morphology, and matches the stimulation intensity, stimulation type, and stimulation time in the cell morphology evolution model. Then, combined with the current location distribution of cells at each differentiation stage, and with the shortest differentiation time as the target, it processes and obtains the stimulation intensity and stimulation type corresponding to each region within the preset culture range at different times. The time and the corresponding region range, stimulation intensity, and stimulation type are used to generate stimulation instructions, which are then sent to the electrical stimulation module. After receiving the stimulation command, the electrical stimulation module processes the area range, stimulation type, and stimulation intensity into operating parameters of the electrical stimulation component, and then applies current to the electrical stimulation component according to the operating parameters, which include current magnitude, time, and location.

2. The direct current stimulation stem cell culture device according to claim 1, characterized in that: The morphological features include cell nuclear features, cell protrusion features, cell outline boundary features and size features, and the arrangement features include cell density distribution features, cell arrangement direction features and key intercellular structural features. The analysis module calculates the two-dimensional length and width of the cell and the total area of ​​the cell based on the size information of the macro image and the cell outline boundary coordinates, and uses the two-dimensional length and width of the cell and the total area of ​​the cell as size features. The cell nucleus outline is segmented to calculate the cell nucleus area, which is then compared with the total cell area to obtain the nucleocytoplasmic ratio. The nucleocytoplasmic ratio and the cell nucleus area are used as cell nucleus features. Extract the pixel set of cell protrusion contours, calculate the protrusion length, width and inflection points to determine the protrusion shape and size, and then count the number of protrusions. Use the protrusion shape, size and number as cell protrusion features. The contour boundary regularity is quantified by the ratio of the actual contour length to the perimeter of the smallest circumscribed polygon, and the contour boundary regularity is used as a feature of the cell contour boundary. During the acquisition of key intercellular structural features, the analysis module combines the identified cell protrusion features and analyzes the extension direction, overlapping area, and connection point information of adjacent cell protrusions through protrusion shape analysis to determine whether key intercellular structures are formed.

3. The direct current stimulation stem cell culture device according to claim 1, characterized in that: The electrical stimulation component includes parallel electrodes and a DC adapter component. The inner side of the culture flask is provided with a recessed culture area. The electrodes are fixedly connected to the culture area, and the parallel structure of the electrodes forms a flushing channel with the mouth of the culture flask. The DC line of the DC adapter component passes through the culture area and is fixedly connected to the culture area. The analysis module uses the direction from the center of the culture range to the mouth of the culture bottle as the reference direction, establishes a reference coordinate system based on the center of the culture range and the reference direction, and marks the position information of each area of ​​the culture range in the reference coordinate system; the stimulation types include pulse stimulation and DC stimulation; The electrical stimulation module stores a parameter matching mapping table containing stimulation intensity, stimulation type, and the magnitude of the current and voltage applied to the electrical stimulation component. The module matches the received stimulation intensity and type against this mapping table to determine the electrical stimulation component and the magnitude of the applied current or voltage. It then adjusts the timing of the applied current to the electrical stimulation component based on the stimulation duration, using this timing, magnitude, and voltage as operating parameters. The electrical stimulation module is electrically connected to the electrodes and DC adapter, and is used to control the timing, magnitude, and voltage of the applied current to the electrodes and DC adapter according to the operating parameters.

4. The direct current stimulation stem cell culture device according to claim 3, characterized in that: A metal sheet is embedded in the side of the culture flask in the culture area. The metal sheet is embedded in the inner wall of the culture flask, and the outer surface of the metal sheet is exposed to the air inside the culture flask. The metal sheet is electrically connected to the power supply through a current regulating device to form a metal sheet electrode, which is an electrode in the electrical stimulation component. A fluorescent fiber optic temperature measuring device is provided on the outside of the culture flask in the area corresponding to the metal sheet. The fluorescent fiber optic temperature measuring device is used to detect the temperature of the area in contact with the metal sheet from the outside of the culture flask. The fluorescent fiber optic temperature measuring device sends the real-time temperature signal to the acquisition module, which then converts the temperature signal into temperature data and transmits it to the analysis module in real time.

5. The direct current stimulation stem cell culture device according to claim 4, characterized in that: The system includes multiple metal sheets distributed in a dot matrix pattern within the culture area. A fluorescent fiber optic thermometer stores a unique identifier for each probe. The fluorescent fiber optic thermometer associates the acquired temperature signal with the probe identifier and sends it to the acquisition module. The acquisition module stores a mapping table between the position information of the metal sheets and the probe identifiers. The acquisition module combines the temperature signals received at the same time with the probe identifiers and their positions in the mapping table to obtain temperature distribution data within the culture area, and then sends this temperature distribution data to the analysis module.

6. The direct current stimulation stem cell culture device according to any one of claims 3, characterized in that: It includes multiple DC adapter components, and the DC wires of the multiple DC adapter components are distributed in a network within the culture area. The unique identification number of each DC adapter component is associated with the location information of the DC wire within the culture area. The electrical stimulation module provides constant DC to only one DC adapter component at a time. The electrical stimulation module matches the DC adapter component to the location of the region range in the culture area according to the stimulation command, and the stimulation intensity matches the current magnitude. The electrical stimulation module stores the stimulation intensity and corresponding stimulation range of each electrical stimulation component under different operating parameters. The electrical stimulation module first obtains the identification number of the electrical stimulation component by matching the stimulation type; the electrical stimulation module matches the identification number of the electrical stimulation component with the region range and the stimulation range, and adjusts the operating parameters of the electrical stimulation component corresponding to the identification number according to the stimulation intensity.

7. The direct current stimulation stem cell culture device according to claim 5, characterized in that: The analysis module stores a temperature mapping table of differentiation stage, differentiation direction and culture temperature range. Based on the current differentiation stage and the differentiation direction corresponding to the preset target morphology, the culture temperature range is matched in the temperature mapping table as a reference temperature range. The reference temperature range is combined with the current cell position distribution of each differentiation stage to obtain the position distribution of the reference temperature range as a reference temperature distribution. The reference temperature distribution and temperature distribution data are sent to the electrical stimulation module. The parameter matching mapping table also stores the heat generation data of each electromagnetic device under different operating parameters; the electrical stimulation module calculates the total heat generation data in each culture area by superimposing the current operating status of each electromagnetic component, compares the total heat generation data with the temperature distribution data to obtain the actual heat dissipation data, compares the reference temperature distribution with the temperature distribution data to obtain the temperature adjustment range, and combines the temperature adjustment range, heat generation data and actual heat dissipation data to adjust the operating parameters of each electrical stimulation component without changing the stimulation range and stimulation intensity.

8. The direct current stimulation stem cell culture device according to any one of claims 1-7, characterized in that: The analysis module calls the built-in cell morphology evolution model, inputs the accumulated environmental stimulus data and the current cell morphology and arrangement characteristics data. The environmental stimulus data includes temperature distribution data, electrical stimulation parameters and nutrient supply related data, and generates the cell differentiation stage prediction results. Based on the inference results, the analysis module generates a sampling distance adjustment command and transmits it to the sampling module. When the cell arrangement characteristics cannot confirm the inference results of the differentiation stage, the sampling module is controlled to shorten the sampling distance for acquiring macro images and increase the sampling frequency. When the cell arrangement characteristics match the inference results, the sampling module is controlled to maintain a preset normal distance and normal frequency to acquire macro images. When the analysis module sends the reference temperature distribution and temperature distribution data to the electrical stimulation module, it generates a temperature adjustment command and sends it to the acquisition module. After receiving the temperature adjustment command, if the cell arrangement characteristics cannot confirm the predicted results of the differentiation stage, the acquisition module shortens the acquisition distance of the macro image and uses the conventional frequency as the acquisition frequency; if the cell arrangement characteristics confirm the predicted results of the differentiation stage, the acquisition distance of the macro image is increased and the acquisition frequency is slowed down.

9. The direct current stimulation stem cell culture device according to any one of claims 8, characterized in that: It also includes a culture medium supply circuit, which is electrically connected to the analysis module; The cell morphology evolution model also includes nutrient consumption rate data. When the analysis module matches the stimulus intensity, it acquires the nutrient consumption rate data. When the difference between the reference temperature distribution and the temperature distribution data is greater than the preset temperature adjustment threshold, the area greater than the preset temperature adjustment threshold is taken as the temperature adjustment area. The temperature adjustment command is generated according to the temperature adjustment area and sent to the culture medium supply circuit and the acquisition module. The culture medium circuit lowers or increases the temperature of the culture medium delivered to the higher temperature area and adjusts the culture medium supply volume. After receiving the temperature adjustment command, the acquisition module increases the acquisition frequency of macro images.

10. The direct current stimulation stem cell culture device according to claim 9, characterized in that: The culture medium circuit has local nozzle components above the parallel paths of each electrode. The nozzle components are fixedly connected to the inner wall of the culture bottle at the upper end of the culture area. The culture medium circuit includes a separate path between each local nozzle component and a control valve that controls the volume of culture medium delivered to the local nozzle component in the separate path. The control valve and temperature regulating device are connected to the separate path respectively. Each temperature regulating device and control valve are electrically connected to the analysis module. The analysis module controls the temperature of the culture medium sprayed to each position in the culture area by using the unique identification number of the temperature regulating device and the corresponding position of the local nozzle component. It controls the volume of culture medium sprayed to each position in the culture area by using the control valve number and the corresponding position of the control valve. The analysis module counts the number of temperature adjustments and the corresponding adjustment time based on the time the temperature regulating device is connected to the current.

11. The direct current stimulation stem cell culture device according to claim 9, characterized in that: The analysis module defines regions that conform to the preset target morphology at the current differentiation stage as successful cultivation regions, and the time corresponding to the current differentiation stage conforming to the preset target morphology as successful cultivation time. The system acquires the operating parameters of each electrical stimulation component of the electrical stimulation module during the successful culture period, constructs an electrical stimulation dynamic model of the electrical stimulation type and intensity distribution at each time point within each culture region, acquires temperature distribution data during the successful culture period, constructs a temperature dynamic model, acquires the supply time and location of the culture medium during the successful culture period, constructs a supply dynamic model, acquires the morphological and arrangement characteristics of cells in macro images during the successful culture period, and constructs a morphological dynamic model. The electrical stimulation dynamic model, temperature dynamic model, supply dynamic model, and morphological dynamic model are time-aligned, and the parameter change trends and correlations of each model are analyzed collaboratively. Based on the collaborative analysis results, the analysis module generates parameter write-back adjustment instructions, writes back and optimizes the electrical stimulation type, electrical stimulation time, electrical stimulation intensity, culture temperature range, and nutrient consumption rate data, and associates the optimized parameters with the location information of the successful culture area in the culture area, storing them as a default template in the built-in database; when a new batch of culture is started, the analysis module generates a template call instruction, and directly applies the default template to the corresponding culture area of ​​the culture bottle.

12. The direct current stimulation stem cell culture device according to claim 11, characterized in that: When the analysis module applies the default template in a new batch of culture, it collects electrical stimulation parameters, temperature distribution data, culture medium supply data, and cell morphology and arrangement characteristics data in real time during the actual culture process. It compares these data with the parameters of each dynamic model corresponding to the default template in real time and calculates the parameter deviation value. When the deviation value exceeds the preset threshold, it generates a real-time fine-tuning instruction to dynamically adjust the electrical stimulation type, culture temperature, and culture medium supply volume in the corresponding area. After the new batch of culture is completed, the analysis module will merge the complete data of the actual successfully cultured area with the original default template data, re-perform multi-model collaborative analysis, generate template update instructions, iteratively optimize the parameters and regional correlations of the default template, and store them.

13. The direct current stimulation stem cell culture device according to claim 12, characterized in that: The electrical stimulation module is equipped with an impedance acquisition circuit. While constructing the dynamic model of electrical stimulation, the analysis module generates electrode function switching instructions to control the dot matrix-distributed metal sheet electrodes to alternately switch between current injection electrodes and voltage detection electrodes. During electrode switching, the impedance acquisition circuit acquires multiple sets of impedance signals and transmits them to the analysis module. The analysis module calls the electrical impedance tomography algorithm to invert the conductivity distribution of the culture medium and extracts data of the cell aggregation region. The cell aggregation region data is then integrated into the morphological dynamic model and subjected to four-dimensional collaborative analysis with the electrical stimulation dynamic model, temperature dynamic model, and supply dynamic model. During parameter write-back optimization, the electrical stimulation intensity gradient and culture medium supply density of the corresponding regions are adjusted in a targeted manner based on the data of the cell aggregation regions.

Images