Continuously updating personalized optimal depth of sedation or other optimal physiologic targeting using processed electroencephalogram signals

Abstract

A personalized targetable optimal level of sedation is derived through automated real-time analysis of an electroencephalogram (EEG) entropy / variability metric and a cerebrovascular reactivity metric. Ongoing real-time collection is made of both EEG entropy / variability metric values and physiological measurements. Artifact detection and removal on raw signals precedes and derivation of real-time cerebrovascular reactivity metric values from the physiological measurements, and calculation of a first and second summary metrics of first and second respective sample sets of the EEG entropy / variability metric values and cerebrovascular reactivity metric values collected in a given period of time. More data points are curated over longer windows of time, and for each such window, are binned by the first summary metric. A curve is fitted to the grouped data points, and a point of lowest cerebrovascular reactivity metric value on the curve is identified, whose EEG entropy / variability metric value denotes the personalized targetable optimal level of sedation.

Description

[0001] CONTINUOUSLY U PDATING PERSONALIZED OPTIMAL DEPTH OF SEDATION OR OTHER OPTIMAL PHYSIOLOGIC TARGETING USING PROCESSED ELECTROENCEPHALOGRAM SIGNALS

[0002] FIELD OF THE INVENTION

[0003] The present invention relates generally to medical equipment, and more particularly to patient monitoring equipment for monitoring a patient’s cerebral physiology.

[0004] BACKGROUND

[0005] Sedation administration is a cornerstone of many medical interventions, from anesthesia for procedures and operations, to medically induced comas within the intensive care unit (ICU), aimed at neuroprotection during states of critical illness. However, the pharmacologic agents used for sedation carry significant downsides, including alterations in systemic blood pressure, derangements in cerebral physiology, and long-term impaired neurological function at high doses. To date, titration of sedation dosing is either based on clinical scoring toolkits applied intermittently at the bedside, or limited physiologic monitoring, that does not focus on optimizing sedation depth or neuroprotection. Recent work has established continuous cerebrovascular reactivity monitoring (surrogate measure for cerebral blood flow control) as a critical predictor of patient outcome in both operative and ICU settings. Moreover, with the advent of electroencephalogram (EEG) data collection and high frequency signal analysis, objective sedation depth can be evaluated via processed EEG entropy metrics, such as the bispectral index (BIS). Marrying the two concepts, recent work by the inventorship group of the present application found a unique individual method of comparing objective sedation depth (using processed EEG - BIS) with cerebrovascular reactivity, thus achieving a theoretical personalized targetable optimal level of sedation, termed BlSopt. This BlSopt target is positioned to be the optimal sedated state, where cerebral blood flow regulation is the most intact, thus positioning the patient in the best neuroprotective state during their procedural sedation, avoiding both under and over sedated states. However, this previous work was limited in that it was only able to be derived after the procedural sedation was completed, required significant manual signal management prior to derivation, and could not be derived in a continuously updating fashion at the bedside. All these limitations had prevented the possibility of this metric being adopted into routine care. Thus was posed the challenge to develop a semi-automated method for continuous derivation of an individualized depth of sedation target in humans.

[0006] The prior work (the topic of prior disclosure in 2022), provided provisional derivation of the BlSopt concept. However, this previous pilot work relied on post-hoc derivation (ie. after the patient left care), necessitated extensive manual signal management prior to derivation, and was not able to be derived in a continuously updating fashion. For such a method to be adopted to clinical practice, these prior technological gaps need to be addressed, producing a semi-autonomous / autonomous platform that is user-friendly for future clinical end-users.

[0007] 2, Background

[0008] As mentioned above, sedative administration is a critical aspect of many patient care plans, spanning anesthesia in operative settings, to medically induced coma states, aimed at preservation of neuronal tissuefl— 3]. However, the drugs employed for such therapeutic sedation techniques carry significant unwanted side-effects. Such negative effects include the reduction in general systemic blood pressure, impairment of cerebral pressure-flow physiologic dynamics, and inducing long-term neuro-cognitive dysfunction in the postoperative patients or post-ICU survivorsfl, 2, 4-6], A contributing factor to this is that sedation depth is often titrated using clinical assessment methods with poor temporal resolution and intra- / inter-assessor variability, or without considering cerebral physiologic responses livetime in an individualized manner[6-8].

[0009] However, newer multimodal monitoring methods for patient cerebral physiology has created new and novel ways to make comments on overall outcome and care[9-l l]. Particularly, continuous cerebrovascular reactivity monitoring (a derived surrogate measure for cerebral blood flow control) has garnered great interest, as it is a cerebral physiologic monitoring metric with moderate / strong associations with neurologic and cognitive outcomes in both neurocritical care [12-19] and intra-operative populations[20, 21], As such, cerebrovascular reactivity metrics have subsequently been investigated as a means to derive individualized blood pressure targets in various operative and ICU states.

[0010] Similarly, continuous EEG monitoring in both operative and ICU settings is positioned to assess sedation depth, particularly utilizing processed entropy metrics derived from raw EEG sources. BIS is one of these proposed entropy metrics that has seen widespread adoption within the operating room setting to monitor sedation depth. The problem is that EEG alone provides no information about cerebral pressure-flow dynamics or oxygen / nutrient delivery. Thus, it cannot be utilized as a means to personalize sedation depth or achieve neuroprotection to prevent the consequences of over or under sedation in operative or ICU settings.

[0011] As such, the solution to optimize depth of sedation monitoring and targeting lies in a fusion between methods of EEG and cerebral pressure-flow dynamic monitoring, to produce a patient-specific depth of sedation metric that focuses on subject-specific physiologic responses to sedation administration and overall neuroprotection. The inventorship group of the present application has been involved in the description of methods to target the least deranged (i.e. optimal) cerebrovascular reactivity state, with a goal of reducing the insult burden caused from impaired reactivity [22-27], Of these methods, a recently described optimal depth of sedation (BlSopt) method has emerged as potentially a unique and minimally invasive way to mediate cerebrovascular reactivity, minimize over-sedation, and focus on neuroprotection during therapeutic sedation administration. BlSopt leverages EEG data collection and high frequency analysis to derive objective measures of sedation depth using EEG spectral analysis[4, 28, 29] through the bispectral index (BIS)[30, 31], Using BIS as a measure of sedation, cerebrovascular reactivity is compared to a range of BIS values to identify the most optimal cerebrovascular reactivity and thus an optimal level of sedation[26, 27], This concept, can be leveraged to any processed EEG entropy metric, and is not just applicable to the BIS EEG measure.

[0012] In past work by the inventorship group, BlSopt was found by leveraging, and improving upon, the fundamental concepts from another optimal cerebrovascular targeting method, optimal cerebral prefusion pressure (CPPopt)

[0032] , CPPopt assesses the systemic blood pressure and links it to cerebral autoregulation (using the Lassen curve relationship)[23, 33], In its most common demonstration, cerebral prefusion pressure (CPP), as a surrogate for systemic blood pressure, is linked to the pressure reactivity index (PRx), as a surrogate for cerebrovascular reactivity (where low values of PRx imply more intact cerebral blood flow regulation / control). CPPopt is determined in individual patients by using a 60-second median CPP calculated alongside PRx, then CPP vs PRx is binned into a boxplot. From this boxplot, the binned CPP value with the lowest associated PRx can be determined (see Figure 1) [23, 27],

[0013] BlSopt is focused on the depth of sedation response through comparing BIS and a cerebrovascular reactivity metric (PRx or non-invasively derived near infrared spectroscopy (NIRS) cerebral oxygen index (COx)). The inventors’ past work has found BlSopt is a unique individualized measure, present in both operative and ICU patient cohorts, while appearing to manifest separately from impaired cerebrovascular reactivity induced by systemic blood pressure (ie. confirmed to have independence from CPPopt and MAPopt values)[26, 27], Moreover, in deriving BlSopt measures there appears to exist good correlation between the BlSopt value derived from invasive intracranial pressure (ICP) monitoring and non-invasive NIRS monitoring metrics[26, 27], Figure 2 provides an example of both patients under general anesthesia (Fig 2A) and those who are critically ill in the ICU (Fig 2B). In both examples, one can see that both states of under- and over-sedation can lead to impaired cerebral blood flow control (worsening cerebrovascular reactivity), exposing the patient to pressure passive states that could lead to stroke, swelling or hemorrhage (and thus neuronal damage and long-term neurological complications). The BlSopt value is the proposed optimal state to target sedation depth (titrating sedation to the BIS value), where the BIS value at this target represents the states where cerebral blood flow control / regulation is the most intact, and thus the safest state for the patient. As such, BlSopt is positioned to make substantial changes to how we administer and target sedated states, focusing on safer care provision and neuroprotection.

[0014] The past work was focused on the initial description / derivation of BlSopt, absent of the intricate details and steps required to ensure a viable clinically relevant BlSopt implementation. Specially, the prior works required all data from the procedure or ICU stay to have been collected to derive BlSopt, making it an “after-the-fact” metric that could not be derived in live-time for clinical care provision. Similarly, there were substantial manual signal management tasks required for the ICP, ABP, NIRS and BIS signal sources, prior to any complex derivation of BlSopt.

[0015] Accordingly, there remains need for a more technically evolved and robust solution for practical real-time implementation of BlSoft, which is the aim of the present application and the inventive systems and methods disclosed herein.

[0016] SUMMARY OF THE INVENTION

[0017] According to a first aspect of the invention, there is provided a computer-implemented method of deriving a personalized optimal physiologic target through automated real-time analysis of an electroencephalogram (EEG) entropy and / or variability metric and a cerebrovascular reactivity metric, said method comprising:

[0018] (a) acquiring, in real-time, and on an ongoing basis, raw input data comprising: real-time EEG entropy and / or variability metric values derived from real-time computer-automated analysis of real-time EEG measurements of a patient; one or more other ongoing real-time physiological measurements of the patient from which real-time cerebrovascular reactivity metric values are derivable through other-real time computer-automated analysis;

[0019] (b) on an ongoing basis, subjecting said raw input data to a real-time computer- automated artifact detection and removal process, thereby deriving cleaned input data void of error-inducing artifacts removed from the raw input data;

[0020] (c) on an ongoing basis, performing said other-real time computer-automated analysis of said one or more other ongoing real-time physiological measurements and deriving said real-time cerebrovascular reactivity metric values therefrom;

[0021] (d) on an ongoing basis, calculating a first summary metric of a sample set of said realtime EEG entropy and / or variability metric values collected in a given window of time and a second summary metric of a sample set of said real-time cerebrovascular reactivity metric values also collected in said given window of time, and saving said first and second summary metrics for said given window of time as a data point;

[0022] (d) curating a plurality of said data points over longer windows of time, and for each of said longer windows of time, grouping said data points into bins, grouped by values of the first summary metric across said longer window of time, and saving said grouped data points as a respective data set;

[0023] (e) for each of said respective data sets, calculating a fitted curve to the grouped data points of said data set, and identifying a lowest point on said fitted curve characterized by a lowest cerebrovascular reactivity metric value, and identifying an EEG entropy and / or variability metric value of said lowest point as a respective approximation of said personalized optimal physiologic target;

[0024] (f) on an ongoing basis, outputting an updated value of said personalized optimal physiologic target that has been derived based, at least in part, on at least one respective approximation from at least one lapsed one of said longer windows of time.

[0025] In at least some embodiments, the personalized optimal physiologic target is a targetable optimal level of sedation, and the EEG measurements of the patient service as an indicator of objective sedation depth of said patient.

[0026] In some embodiments, said EEG entropy metric is an EEG bispectral index (BIS) and said real-time EEG entropy metric values are BIS values received from a BIS monitor that processes the EEG measurements of the patient.

[0027] In some embodiments, said real-time cerebrovascular reactivity metric is an intracranial pressure (ICP) derived cerebrovascular reactivity metric.

[0028] In some embodiments, said real-time cerebrovascular reactivity metric is selected from a group consisting of a pressure reactivity index (PRx), a pulse amplitude index (PAx), and a correlation between pulse amplitude of ICP and cerebral perfusion pressure (RAC).

[0029] In some embodiments, said real-time cerebrovascular reactivity metric is a near infrared spectroscopy (NIRS) derived cerebrovascular reactivity metric.

[0030] In some embodiments, said NIRS derived cerebrovascular reactivity metric is selected from a group consisting of a cerebral oxygen index, an oxyhemoglobin (HbO) index (HbOx), a deoxyhemoglobin (Hb) index (Hbx), a tissue oxygen saturation (TOI) index (TOx), and a total hemoglobin (THI) index (THx). In some embodiments, said other one or more physiological measurements comprises at least two different physiological measurements and other-real time computer-automated analysis comprises a Pearson correlation between respective series of consecutive mean values for each of said at least two different physiological measurements.

[0031] At least some preferred embodiments comprise receiving said other one or more physiological measurements from one or more non-invasive patient monitors.

[0032] In at least some preferred embodiments, step (b) comprises excluding, from the cleaned data, real-time physiological measurements falling outside predetermined thresholds that denote opposite ends of a reasonable measurement range.

[0033] In some embodiments, step (b) comprises performing detection of pulse wave forms in at least one of said one or more other ongoing real-time physiological measurements, and excluding from said cleaned input data any measurement signals that both exceed a predetermined length and are void of any said pulse wave forms.

[0034] In some embodiments, step (b) comprises monitoring electromyography (EMG) signals for instances of extreme magnitude, and for any EEG entropy and / or variability metric values coinciding with such instances, excluding said any EEG entropy and / or variability values from said cleaned input data.

[0035] In some embodiments, step (e) further comprises calculating a curve- strength factor based on difference of magnitude between a lowest cerebrovascular reactivity metric value among the grouped data points of the data set, and a higher one of two cerebrovascular reactivity metric values each corresponding to a respective end of the fitted curve.

[0036] Such embodiments preferably use said curve-strength factor for at least one of the following:

[0037] (i) weighted determination of the updated value in step (e) based on a plurality of respective approximations from a lapsed plurality of said longer windows of time, which are weighted based at least partly on said curve- strength factor; and / or

[0038] (ii) output of said curve strength factor, or other readable curve strength indicator derived therefrom, with the updated value in step (f) to give added context to said updated value.

[0039] In some embodiment, step (e) further comprises calculating a lowest-point location factor characterizing a locality of the lowest point along the fitted curve relative to a quantity of bins into which the data points were grouped.

[0040] Such embodiments preferably use said lowest-point location factor for at least one of the following:

[0041] (i) weighted determination of the updated value in step (e) based on a plurality of respective approximations from a lapsed plurality of said longer windows of time, which are weighted based at least partly on said lowest-point location factor; and / or

[0042] (ii) output of said lowest-point location factor, or other readable location-pertinent indicator derived therefrom, with the updated value in step (f) to give added context to said updated value.

[0043] In at least some embodiments, step (f) comprises displaying said updated value, or an equivalent readable indicator of said personalized targetable optimal level of sedation, in a local environment containing said patient and / or one or more medical practitioners seeing to said patient.

[0044] In some instances, said local environment is an intensive care unit.

[0045] In some instances, said local environment is an operating room.

[0046] Some embodiments further comprise (g) sending control signals to an intravenous sedation titration device to control titration of sedation dosing thereby within a closed loop targeting the updated value of said personalized targetable optimal level of sedation.

[0047] In at least some embodiments, said summary metrics are median values of said sample sets.

[0048] In at least some embodiments, said one or more other ongoing real-time physiological measurements comprise at least two real-time physiological measurements.

[0049] According to a second aspect of the invention, there is provided an electronic device for deriving a personalized optimal physiologic target through automated real-time analysis of an electroencephalogram (EEG) entropy metric and a cerebrovascular reactivity metric, said device comprising one or more processors and non-transitory computer readable memory coupled thereto, in which there are stored statements and instructions that are executable by said one or more processors and are configured to, when executed, perform at least steps (a) through (f) of the method recited in the preceding aspect of the invention.

[0050] In at least some embodiments, the device comprises an electronic display, and the statements and instructions are further configured to display the updated value of said personalized targetable optimal level of sedation, or an equivalent readable indicator of said personalized targetable optimal level of sedation, on said electronic display.

[0051] In at least some embodiments, the device is provided or used in combination with an intravenous titration device connected thereto, wherein the statements and instructions are configured to send control signals to the intravenous titration device to control titration of dosing thereby in a closed feedback loop targeting the updated value of said personalized optimal physiologic target.

[0052] According to a third aspect, the invention also extends to a non-transitory computer readable memory having stored therein statements and instructions executable by one or more processors to perform at least steps (a) through (f) of the first aspect of the invention.

[0053] BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Preferred embodiments of the invention will now be described in conjunction with the accompanying drawings in which:

[0055] Figure l is a boxplot of cerebral prefusion pressure (CPP) vs. pressure reactivity index (PRx), as respective surrogates for systemic blood pressure and cerebrovascular reactivity, from which an optimal CPP (CPPopt) is identifiable as the binned CPP value with the lowest associated PRx, as is known in the art.

[0056] Figure 2 shows a boxplot (A) of bispectral index (BIS) vs. cerebral oxygen index (COx) from an elective spinal surgery patient, and a boxplot (B) of BIS vs. pressure reactivity index (PRx) for a traumatic brain injury patient, from each of which an optimal BIS (BlSopt) is identifiable in accordance with the present invention. Figure 3 A is a combined block and process diagram of one embodiment of an inventive system and method of the present invention for monitoring BlSopt in real time to quantity a patient’s depth of sedation and prevent neurologically detrimental oversedation through human or automated adjustment of titrated sedation dosing.

[0057] Figure 3B is a basic block diagram of a computing device for executing automated steps of the process of Figure 3 A.

[0058] Figure 4 illustrates examples of errors that may occur in raw signals from patient monitors that measure input patient variables in the Figure 3 process, including pinching of a blood pressure line (A), an error or failure in device sensor (B), an increase due to drug infusions (C), and a loss of signal (D).

[0059] Figure 5 shows plots of continuous PRx and BIS derived as useful input to real time calculation of BlSopt.

[0060] Figure 6 is a boxplot illustration derivation of BlSopt, showing plotted datapoints of BIS vs. PRx, binning of those datapoints and plotting of a fitted curve to the binned datapoints, own which curve the lowest point is the determined BlSopt.

[0061] Figure 7 shows multiple boxplots of equivalent type to Figure 6, but showing fitted curves of differing character, which are assessed for strength and weighted differently in the Figure 3 process according to assessed weakness (A & B) or strength (C & D).

[0062] Figure 8 schematically shows displayed final output from the system of Figure 3, including a BlSopt value indicative of sedation strength, and associated scores for imparting addition context to the displayed BlSopt value.

[0063] DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS

[0064] Detailed herein in an enabling and reproducible manner are the steps required to create a clinically relevant BlSopt value, and practical and automated implementation of such method in a clinical setting. This involves the automated manipulation and cleaning of raw signal artifacts, such that the data is primed for future calculations. Second, the cleaned data will be processed to derive both depth of sedation (BIS) and cerebrovascular reactivity (PRx / COx) in a continuous fashion. Finally, the optimal calculation of BlSopt is completed, which will outline the value, confidence in the value, accuracy and other aspects surrounding the optimal calculation of BlSopt. Such steps need to be evaluated in a research and clinical setting to ensure the real-time functionality of our system. Finally, the final product requires all of these steps stored in a software package that is user friendly and ready for implementation leveraging data from existing monitoring devices.

[0065] Based on the above-mentioned challenges left unanswered by the prior work, the present inventors sought to address these and develop a more robust and efficient automated system. This includes the software aspects associated with an implementable system, and a hardware product for practical application of the inventive concept, in totality thereby demonstrating the invention’s viability in a clinical setting. The problems that were addressed will be outlined in detail in the paragraphs below, followed by details of the inventive solutions to these problems / challenges that have been pursued and achieved through sequential technological advances, generating a novel and new system-based solution to the BlSopt derivation from physiological data. Included among the present disclosure are unique aspects surrounding data management, curation, derivation and demonstration of BlSopt, such that a real-time clinically relevant value of BlSopt can be continuously found, in a completely novel fashion unachieved by the prior art.

[0066] Figure 3 is a combined block diagram and process flowchart illustrating both a system and method of one preferred embodiment of the present invention. The system 10 comprises a group of patient monitors 12A-12C respectively connected to a group of sensors 14A-14C operably placed on the patient 16, among which one of the monitors 12A performs real-time computer-automated analysis of real-time EEG measurements of the patient 16 from EEG sensor 14A to derive real-time EEG entropy and / or variability metric values, and in the illustrated embodiment is a BIS monitor deriving real-time BIS values. The one or more other patient monitors 12B, 12C take ongoing real-time physiological measurements of the patient 16 from their respective sensors 14B, 14C, and in the illustrated example are embodied by an intracranial pressure (ICP) monitor) 12B connected to an ICP sensor 14B to measure the patient’s ICP, and an arterial blood pressure (ABP) monitor 12C connected to an ABP monitor 14C to measure the patient’s ABP. A computing device 18 is connected to the monitors 12A- 12C to receive and process the raw BIS, ICP and ABP data signals therefrom. As schematically illustrated in Figure 3B, the computing device 18 comprises one or more processors 20, random access memory (RAM) 22 and non-volatile memory 24, both of which memories are non-transitory computer readable memory for reading and writing of data thereto and therefrom by the processor(s) 20, and a visual display 26. All such hardware componentry of the computing device 18 are interconnected, directly or indirectly, by one or more busses 28 to enable any and all herein described cooperative functionality between these components. The non-volatile memory 24 has stored therein inventive software algorithms for transforming the raw signal data from the monitors 12A-12C into the ultimately calculated and display BlSopt information. The software is composed of executable statements and instructions for execution by the processor(s) 18 to perform any all algorithms, processes, routines, tasks, and steps described herein, except for any of those that may be explicitly described as being performed by another means or actor. In Figure 3, the software is schematically denoted by functional blocks 24A-24C of a flow diagram representative of the algorithmic workflow of the software, though it will be appreciated that the software, while being implementable in modules, need not necessarily be implemented via modules of matching description to individuated functional blocks of the schematic illustration.

[0067] Clinical data management in an automated fashion needs to be completed such that the BlSopt derivation can be implemented in real-time. The hardware aspects of data collection can be completed using pre-existing techniques and hardware, and thus a bulk of the following disclosure on the steps and algorithms used to curate the data for analysis, albeit without detracting from the ultimate practical application of those aspects into a patentable invention. On the data handling side, there was a need to focus on two primary aspects of data manipulation (aka - Sub-Challenges): 1. Raw Signal Artifact Detection / Removal and, 2. Derivation of Fundamental Values.

[0068] The raw collected signals (in this case BIS, arterial blood pressure; ABP, intracranial blood pressure; ICP, and brain tissue oxygen saturation; rSO?) required for the calculation of BlSopt need to be automatically processed for artifacts that occur naturally during aspects of care. Figure 4 demonstrates some of these errors, including their form and the resulting issues that occur for their removal. As such this aspect of the inventive solution, implemented at functional block 24A of Figure 3 and detailed further below, is aimed at either removal or interpolation of these artifacts in an automatic fashion.

[0069] With the cleaned data, various values of importance need to be derived from raw clean signals, which derivation is implemented at functional block 24B of Figure 3. Figure 5 is an example of PRx (cerebrovascular reactivity metric; surrogate for cerebral blood flow control) and BIS calculated for the final BlSopt derivation. Again, this needs to be completely automatic and prepared in a fashion for data analysis. It should also be noted that for a minimally / non-invasive implementation of the BlSopt methodology, the NIRS-based cerebral oximetry index (COx; derived using rSCh and ABP) can be used as a cerebrovascular reactivity.

[0070] It is worth noting that there are various continuous cerebrovascular reactivity indices that can be derived from one of a few methods. Almost all methods leverage two signal sources, including a signal for driving pressure (either ABP or cerebral perfusion pressure (CPP), and a signal for cerebral pulsatile blood volume (through ICP or NIRS devices) or blood flow (through transcranial Doppler (TCD), parenchymal thermal diffuse flowmetry (TDF), or brain tissue oxygen monitoring (PbtCh). These input / output signals can be used to approximate cerebrovascular reactivity through one of a few methods: A. time-domain (leveraging Pearson or similar correlation coefficients), B. frequency domain (using transfer function or similar impulse response analysis), or C. wavelet coherence / phase-shift analysis (often employ cosine functions applied to phase-shift values), all continuously calculated using sliding window approaches. In the present disclosure, in the interest of brevity, primary focus is given to one example of an ICP -based index (ie. PRx) and one example of a NIRS-based index (ie. COx). However, the same methodology for an optimal sedation metric derivation can be applied to any / all continuously derived cerebrovascular reactivity measures. A summary of such measure can be found in Table 1 below: Table 1: Examples of Continuously Derived Cerebrovascular Reactivity Indices - Time- Domain, Frequency-Domain, and Wavelet Methods

[0071] Legend: ABP = arterial blood pressure, AMP = pulse amplitude of ICP (fundamental amplitude), CPP = cerebral perfusion pressure, Dx / Dx-a = diastolic flow index derived from TCD, FVd = diastolic cerebral blood flow velocity, FVm = mean / median cerebral blood flow velocity, FVs = systolic cerebral blood flow velocity, Hb = deoxyhemoglobin, Hbx / Hbx-a = hemoglobin index, HbDiff= difference between oxy- and deoxyhemoglobin, HbDiffx / HbDiffx- a = hemoglobin difference index, HbO = oxyhemoglobin, HbOx / HbOx-a = oxyhemoglobin index, HbTot = total hemoglobin, HbTotx / HbTotx-a = total hemoglobin reactivity index, LDF = laser Doppler flowmetry, LDFx / LDFx-a = laser Doppler flow index, Mx / Mx-a = mean flow index derived from TCD, NIRS = near infrared spectroscopy, ORx / ORx-a = oxygen reactivity index, PAx = pulse amplitude index, PbtCh = brain tissue oxygen saturation, PRx = pressure reactivity index, RAC = correlation between AMP and CPP, rSCh = regional cerebral oxygen saturation, Sx / Sx-a = systolic flow index derived from TCD, TF-ARI = transfer function autoregulatory index, TDF = thermal diffusion flowmetry, TDFx / TDF-a = thermal diffusion index, THI = total hemoglobin index, THx / THx-a = total hemoglobin saturation index, TOI = tissue oxygen index, TOx / TOx-a = tissue oxygen reactivity index. Note: “w ” prefix to all indices denotes “wavelet” version.

[0072] From the curated data suitably processed to overcome the data management problem, the full continuous BlSopt derivation is vital for the optimal performance of the targeted sedation approach in a clinical environment. This will not only include the need for the continuously derived value of BlSopt, but details around the resulting calculation behind the value, the overall confidence and other aspects associated with its derivation. This is due to the fact that, for clinical adoption, BlSopt ideally needs to provide the clinician with the overall confidence in the value. Figure 6 demonstrates BlSopt and the various factors that will be associated with it.

[0073] The whole data curation and processing aspect of this system will need to be implemented within a data format ready for clinical and research use. Meaning, it cannot be reasonably expected for a clinician to perform complicated biomedical engineering and signal processing tasks at the bedside in order to derive and utilize BlSopt. Thus, a full packaged system will need to complete the outlined task in Figure 3 (Key Problems #1 and 2).

[0074] For commercial viability, the final product must be fully ready for implementation within a clinical environment. As such the system must collect raw physiological signals for BlSopt derivation including: depth of sedation and the raw signals to derive cerebrovascular reactivity. This raw data will be curated and processed for BlSopt calculations in real-time and display the finalized results. Finally, this information needs to be wrapped up into a user-friendly software / hardware package, that the average person can implement for their own personal data collection.

[0075] In the paragraphs below, the steps to semi-automate / automate data management pipelines and continuously derive BlSopt are detailed, with particular focus on inventive solutions to the above referenced problems and sub-challenges. Some embodiments may leverage open-source software as much as possible for the creation and optimization of the data and BlSopt calculations. On the hardware side of things, commercially available hardware products that may populate a system to curate and manipulate the data, without detraction from embodiment of the inventive concepts into a practical application of patentable character.

[0076] On the data management side of things, the data manipulation involves the basic methods for the curation and preparation of the raw data for BlSopt calculations. This needs to be automatic without any direct involvement of a medical professional, thus the two areas of data processing: 1. artifact detection / removal and 2. derivation of fundamental values. Prototyped embodiments employed Python and R Statistical Computing.

[0077] Concerning the sub-challenge of Raw Signal Artifact Detection / Removal, Figure 4 demonstrates how there may be many errors across various physiological variables that must be curated and accounted for in BlSopt derivation. The inventors endeavored to build an automated coding pipeline that detects and removes these artifacts using common and some complex principles. The primary method implemented at functional block 24A of Figure 3 is thresholding values of ICP and ABP that fall within reasonable ranges, for example ICP / ABP >0mmHg, ABP <300mmHg and ICP <100mmHg. BIS and rSCh also can be identified as having values between 0-100, with values that drop directly to 0 as erroneous. Next, at the schematically labelled “erroneous signal handling” step of functional block 24A, with the adoption, in at least some embodiments, of the Pan-Tompkins method and / or continuous wavelet transform methods for viable pulse-waveform signal detection, the ICP / ABP / rSCh waveforms can be evaluated for the presence of cardiac, respiratory and other known normal physiologic related pulse wave form morphology and frequency band power, to confirm viability

[0034] . In signals where this pulse waveform morphology was not seen for more than 10 seconds the signal can be identified as erroneous over this range (Figure 4 Panel B), and thus discarded at the erroneous signal handling step of functional block 24A. Finally, given the nature of the collected BIS signals, there is electromyography (EMG) information that can be used in preferred embodiments to evaluate the effectiveness of BIS (processed EEG). Extreme EMG signals can be determined through massive increases in the overall amplitude, noting that this would interfere and disrupt BIS and thus will be removed at this erroneous signal handling step. In some embodiments, these times are marked as erroneous and filled with NA (not available) values, thus the BlSopt method does not use them for calculations. This is done automatically in real-time, without the oversight of a medical professional and thus is ready for the next steps of our method. Subsequently, after both “thresholding” and “ erroneous signal handling” steps are completed for raw ICP / ABP / rSCh / BIS signal sources, the “averaging” step of functional block 24A is then applied, decimating all cleaned signal sources using a nonoverlapping moving average filter (ie. low band-pass filter) to focus on data resolution ranges affiliation with cerebrovascular slow-waves.

[0078] Turning to the sub-challenge of the Derivation of Fundamental Values from the cleaned data void of signal artifacts, the key variables then need to be found for BlSopt calculation at functional block 24B of Figure 3. In some embodiments of the inventive methodology, a median BIS found over a minute is compared to a PRx value found over a minute. In some embodiments, PRx as a surrogate for cerebrovascular reactivity is found using the Pearson correlation between 30 consecutive 10-s mean values of ICP and ABP physiology, updated at some regular interval, for example every minute[9, 14, 15, 35, 36], The calculated values of BIS and PRx are updated at every consecutive interval (e.g. minute) of recording and fed into the BlSopt method for real-time data analysis.

[0079] For a minimally / non-invasive implementation of BlSopt, rSCh can be used in the derivation of the cerebrovascular reactivity measure of COx. Like PRx, the current minimally invasive method finds a median BIS over a minute and is compared to a COx value found over a minute. COx as a surrogate for cerebrovascular reactivity is found, for example using the Pearson correlation between 30 consecutive 10-s mean values of rSO2 and ABP physiology, updated every minute[l 1, 37-40], The calculated values of BIS and COx are updated at every consecutive interval (e.g. minute) of recording and fed into the BlSopt method for real-time data analysis. Depending on the initial collected data whether ICP, rSO2 or both, will dictate which cerebrovascular reactivity measure is used. Though for the final BlSopt calculation only one cerebrovascular reactivity value is needed.

[0080] Turning from data management to Continuous BlSopt Derivation and Demonstration, implementation of which is made at functional block 24C of Figure 3, the presently disclosed and inventive embodiment of real-time BlSopt implementation builds on the past aspects of CPPopt methodology [10, 15, 41-43], but with important novel algorithmic modifications to address curve fit and yield issues specific to the BIS vs PRx / COx relationship. In one embodiment, the first step is to find the median value of BIS and PRx / COx values over a prescribed period, for example 5-minute window. With the curated and cleaned minute-by- minute data of PRx / COx and BIS, the data is then binned into 5au of BIS across the range of BIS values over 1 / 2 / 4 / 6 / 8 hours window of time and compared to PRx / COx to give a Boxplot for each window of time as shown in Figure 2 / 6 / 7. Like past CPPopt methods to improve stability, confidence and yield, a formulaic weight is applied to these windows, such that the best performance window would have a more significant impact on the BlSopt values (Equation 1). CPPopt location * Curve Strength (1)

[0081] Equation 1 BlSopt Location, location of the BlSopt value in the curve (0 / 0.75 / 1); Curve Strength, the strength of the curve depending on range of PRx / COx (weak to very strong 0.5 / 0.75 / 1); jullfiterror, R2 heuristic from the fitted data(values between 0.2 - 1); Weight, final factor for weight adjustment; windowlength, is the time length of the window used (time window limitations, usually 2 to 8 hours).

[0082] Thus, in the present embodiment, for each window of time (e.g. 1 / 2 / 4 / 6 / 8 hours) the following steps and calculations are performed by the software at functional block 24C of the illustrated workflow to find a weighted adjustment by which to calculate a final BlSopt value in this preferred embodiment of the BlSopt method:

[0083] A. BlSopt location has been noted in the past to result in drastically varying BlSopt values over a short period of time (increasing volatility). Figure 7A / B demonstrate examples of an ascending / descending curve BlSopt fitted values, and all BlSopt values derived with this relationship are removed. Next, it is noted that some BlSopt derived values have a BlSopt in the middle of the BIS bins but close to the tail ends of the data (low BIS variation or values), and though a BlSopt with a U-shaped curve could be found, it is difficult to determine if the derived BlSopt is well balanced. To this end, a distinction is made between the U-shaped curves with many BIS bins (>5bins) and the BlSopt value 2 bins away from the max / min BIS bins. Thus, BlSopt location are assigned different weighting values (e.g. between 0 and 1), for example a low locationweighting value (e.g. 0) in the case of ascending / descending curve (Fig 7A / B), an intermediate location-weighting value (e.g. 0.75) for low CPP bin count or BlSopt skewedness (Fig 7C), or a high location-weighting value of (e.g. 1) for a good number of BIS bins and strong location (Fig 7D).

[0084] B. Curve Strength is used herein to refer to the overall magnitude strength between the lowest BIS binned PRx value and the highest on either end of the curve. The PRx difference between the two bins is found in some instances to be in an upper threshold range (e.g. >0.3) denoting a strong curve that is therefore assigned a high curveweighting value of 1 (strong curve = 1), in other instances to be in a middle threshold range (e.g. 0.2-0.3) denoting a moderate curve that is therefore assigned an intermediate curve-weighting value (e.g. 0.75), or in other instances to be in a low threshold range (e.g. <0.2) denoting a weak curve that assigned a low curve-weighting value (e.g. 0.5). In this way, a magnitude for the BlSopt can be quantified and displayed as contextual feedback on the calculated and displayed BlSopt value.

[0085] The final BlSopt value, having been weight adjusted according to both location and curve strength, is displayed on the visual display 26 of the computing device 18 in real time along with the key aspects like BIS location score (e.g. the applied location-weighting value, or some other location scoring metric derived therefrom), curve strength score (e.g. the applied curve-weighting value, or some other curve strength scoring metric derived therefrom) and overall confidence score.

[0086] Confidence (as displayed in Figure 8) is a demonstration of the overall BlSopt calculated value, taking into account various aspects of the each overall calculated BlSopt window values, the factors that make up each calculated window and its overall consistency between them. The full confidence value is found using each window’s ‘Weight of Window’ (from equation 1) and the stability of ‘Weight of Window’ (ie. is there a high overall Weight of Window for each BlSopt value) and the consistency in the resulting BlSopt values. Equation

[0087] 2 is a summary of the equation with overall higher values resulting in a higher confidence.

[0088] Confidence = A*5(Weights of Windows ) + B * mean(Weights of Windows ) + C *5(BISopt Windows Value) (2)

[0089] Equation 2 BlSopt windows value, value of BlSopt found for each window; Confidence, overall found confidence for the final displayed BlSopt value; Weights of Windows, each found windows individual weighted value; mean(), the mean value;5(), the standard devation for the given values; A,B,C are modifiers.

[0090] Note for ease of understanding and reduction of complicated numbers for the end-user, the final calculated numerical confidence score is transformed in the illustrated embodiment to a non-numerical expression of confidence for final text output (for example High, Moderate and Low), though such transformation is not essential, and may be omitted in other embodiments that instead display the numerical confidence value instead of a descriptive non- numerical expression of confidence.

[0091] Such work marks as a unique demonstration of the optimal cerebrovascular reactivity value and marked improvement over the prior work. In similarly beneficial advancement over the prior work, implementation of a sliding derivation method results in a system that can be used to determine BlSopt automatically, from collected patient data into a continuously updating real time BlSopt value. Display of the real time BlSopt on the visual display 26 to an onsite anesthesiologist 29 responsible for the patient’s sedation informs the anesthesiologist of the patient’s presently targeted depth of sedation, and enables adjustment by the anesthesiologist of the titration device 30 to adjust the titration of sedation dosing of the patient 16 relative to the targeted depth of sedation denoted by the displayed BlSopt, to ensure avoidance of neurologically detrimental oversedation. The illustrated embodiment of Figure

[0092] 3 illustrates optional implementation of a closed feedback loop 32 in which the computing 1 device 18 and the titration device 30 are connected to enable optionally automated control of the titration device 30 by outputted commands from the computing device 18 based on the real time calculated BlSopt.

[0093] Demonstration was also made of the utility of the inventive system that determines BlSopt automatically from collected patient data and outputs an updated real time value Ethics for this data were collected following full approval by the University of Manitoba Health Research Ethics Board (H2017:181, H2017: 188, B2018: 103 and B2019:065) and the Health Sciences Centre Research Impact Committee (R2019:072). All patients had informed consent obtained as part of B2018: 103. Using PRx as a cerebrovascular reactivity measure for BlSopt derivations, in a cohort of 72 traumatic brain injury patients of over 4500 hours of data, BlSopt was derived in a continuously updating fashion. BlSopt could be found at least once in 69 (96%) of the patients, with a BlSopt median value of 44 (34-56) and overall % yield of 67 (55- 87) for continuous derivation. Similarly, using the minimally invasive COx as a cerebrovascular reactivity measure for BlSopt calculation, in a cohort of 65 traumatic brain injury patients of over 3600 hours of data, we were able to derive BlSopt in a continuous and entirely non-invasive fashion. BlSopt could be found at least once in 65 (100%) of the patients, with a BlSopt median value of 44 (36-54) and overall % yield of 71 (62-88) for continuous derivation.

[0094] Prototyping of the software was implemented using R statistical computing for the demonstration of the BlSopt calculations. This included the primary BlSopt calculation and the preceding data manipulation / cleaning. Though prototyping was done on collected raw ICP / ABP / rSCh / BIS data, the data structure allows for this to be seamlessly adapted from real time data sources given the streamlined nature of the code. In a prototyped embodiment, the BlSopt finalized values are displayed on a display screen of the hardware in the manner shown in Figure 8, though the particular organization layout, labels and scoring terminology may vary in other embodiments. The illustrated example displays the BlSopt value, a Bin Location score, a Curve Strength score and a Confidence score, denoting results from the key variables used in the prototyped embodiment, more variables can be added in other embodiments expanding upon the prototyped example. Further refinements that may be undertaken in relation to artifact management and curation to improve the overall accuracy of BlSopt include the evaluation of signal waveforms, time-domain analysis, frequency-domain analysis and emerging knowledge around artifact morphology.

[0095] Work surrounding cerebrovascular reactivity has found that this measure can be limited and highly volatile. This is a limitation to CPPopt as well, thus new methods for determining cerebral autoregulation can explored given that overall, they may offer improved stability to cerebrovascular reactivity. Emerging examples of this include wavelet PRx and the pulse amplitude index, which have both demonstrated an improved performance to PRx[44-47], Thus, efforts can be made to use the most up to date and accurate cerebrovascular reactivity measure. As such, the above-mentioned EEG-based entropy / variability approach (BlSopt provided as a working example) to optimal sedation state targeting can therefore utilize any continuously derived cerebrovascular reactivity metric. Such continuous cerebrovascular reactivity metrics include those derived through Pearson correlation methodology in the timedomain, frequency-domain transfer function methods, or wavelet coherence / phase-shift methods, leveraging sliding window calculations for continuous data stream derivation, (as mentioned elsewhere above)

[0096] In some embodiments, the EEG entropy index of BIS is used to classify an objective measure of depth of sedation. BIS was selected because it is one of the most common and readily available methods, used clinically. However, given that this work is linking depth of sedation and cerebrovascular reactivity, implementation using other objective measures of sedation depth is also possible (other EEG entropy indices other than BIS can also be used to find BlSopt). Moreover, as a clinically relevant value, BIS may not be the best / most robust method for determining sedation state in critical care patients. Efforts may nonetheless continue to improve and modify the limitations of EEG derived entropy indices (like BIS) to improve their accuracy. Such work may leverage more transparent entropy measures (Sample, Apparent, Multi-scale) derived from raw EEG waveforms. Furthermore, such work will go into the assessment of which EEG aspects are associated with depth of sedation and cerebrovascular reactivity. Therefore, the BlSopt method may not use BIS in some future or alternative embodiments, which may instead use some other EEG-based objective measure of sedation depth.

[0097] Implementations of the present invention can also be extended beyond specific application to targeted depth of sedation, and can be similarly applied to derive other personalized optimal physiologic targets likewise updated and outputted in real-time relationship to collected patient measurements from which they are derived. Workable examples of such other applications have been demonstrated to include optimal cerebral prefusion pressure (CPPopt), and optimal arterial blood pressure (ABPopt), the later of which may alternatively be referred to as optimal mean arterial blood pressure (MAPopt). The algorithmic approach is similar to that described of the preceding detailed embodiments for BlSopt - first cleaning the raw data from the patient monitors then evaluating either the minimum between CPP and any one of the listed or derived cerebrovascular reactivity metrics to derive CPPopt, or the minimum between ABP / MAP and any one of the listed or derived cerebrovascular reactivity metrics to derive ABPopt / MAPopt. In similar implementation of a closed feedback loop for at least semi-automated patient dosing embodiments, an intravenous titrator can once again be connected to the computing device for real-time feedback control thereby, but instead of titrating sedation agents, the titrator instead performs titration of vasopressor medications to manipulate the monitored pressures (ie. CPP or ABP / MAP) to the optimal target range.

[0098] In summary, the developed BlSopt concept and algorithms are positioned as a unique method for individualized depth of sedation targeting, focused on neuroprotection, which can be leveraged in multiple operative and ICU settings where therapeutic sedation is administered. The inventive system takes collected raw physiological data and manages, curates, and implements it in the continuous derivation of BlSopt in a commercially and clinically viable system for continuous individualized depth of sedation targeting, optimizing the relationship between cerebral pressure-flow and EEG based physiologies.

[0099] Since various modifications can be made in the invention as herein above described, and many apparently widely different embodiments of same made, it is intended that all matter contained in the accompanying specification shall be interpreted as illustrative only and not in a limiting sense. 8. References

[0100] 1. Flower O, Hellings S (2012) Sedation in Traumatic Brain Injury. Emerg Med Int 2012:. https : / / doi. org / 10.1155 / 2012 / 637171

[0101] 2. Menon DK, Young Y, Tew DN, Bacon PJ (1994) New horizons in ICU sedation: exploring non-sedative effects of ICU sedation. Clin Intensive Care 5:22-26

[0102] 3. Oddo M, Crippa IA, Mehta S, et al (2016) Optimizing sedation in patients with acute brain injury. Crit Care 20:. https: / / doi.org / 10.1186 / sl3054-016-1294-5

[0103] 4. Froese L, Dian J, Gomez A, et al (2021) Association Between Processed Electroencephalogram-Based Objectively Measured Depth of Sedation and Cerebrovascular Response: A Systematic Scoping Overview of the Human and Animal Literature. Frontiers in Neurology 12: 1409. https: / / doi.org / 10.3389 / fineur.2021.692207

[0104] 5. SRLF Trial Group (2018) Impact of oversedation prevention in ventilated critically ill patients: a randomized trial-the AWARE study. Ann Intensive Care 8:93. https: / / doi.org / 10.1186 / sl3613-018-0425-3

[0105] 6. Ely EW, Truman B, Shintani A, et al (2003) Monitoring Sedation Status Over Time in ICU Patients: Reliability and Validity of the Richmond Agitation- Sedation Scale (RASS). JAMA 289:2983. https: / / doi.org / 10.1001 / jama.289.22.2983

[0106] 7. Arevalo JJ, Brinkkemper T, van der Heide A, et al (2012) Palliative sedation: reliability and validity of sedation scales. J Pain Symptom Manage 44:704-714. https: / / doi.Org / 10.1016 / j.jpainsymman.2011. l l.010

[0107] 8. Hogg LH, Bobek MB, Mion LC, et al (2001) Interrater reliability of 2 sedation scales in a medical intensive care unit: a preliminary report. Am J Crit Care 10:79-83

[0108] 9. Donnelly J, Czosnyka M, Adams H, et al (2019) Twenty -Five Years of Intracranial Pressure Monitoring After Severe Traumatic Brain Injury: A Retrospective, Single-Center Analysis. Neurosurgery 85:E75-E82. https: / / doi.org / 10.1093 / neuros / nyy468 10. Depreitere B, Giiiza F, Van den Berghe G, et al (2014) Pressure autoregulation monitoring and cerebral perfusion pressure target recommendation in patients with severe traumatic brain injury based on minute-by-minute monitoring data. J Neurosurg 120:1451-1457. https: / / doi.Org / 10.3171 / 2014.3.JNS131500

[0109] 11. Zeiler FA, Donnelly J, Menon DK, et al (2017) Continuous Autoregulatory Indices Derived from Multi-Modal Monitoring: Each One Is Not Like the Other. Journal of Neurotrauma 34:3070-3080. https: / / doi.org / 10.1089 / neu.2017.5129

[0110] 12. Laurikkala J, Aneman A, Peng A, et al (2021) Association of deranged cerebrovascular reactivity with brain injury following cardiac arrest: a post-hoc analysis of the COMACARE trial. Critical Care 25:350. https: / / doi.org / 10.1186 / sl3054-021-03764-6

[0111] 13. Budohoski KP, Czosnyka M, de Riva N, et al (2012) The Relationship Between Cerebral Blood Flow Autoregulation and Cerebrovascular Pressure Reactivity After Traumatic Brain Injury: Neurosurgery 71 :652-661. https: / / doi.org / 10.1227 / NEU.0b013e318260febl

[0112] 14. Depreitere B, Citerio G, Smith M, et al (2021) Cerebrovascular Autoregulation Monitoring in the Management of Adult Severe Traumatic Brain Injury: A Delphi Consensus of Clinicians. Neurocrit Care, https: / / doi.org / 10.1007 / sl2028-020-01185-x

[0113] 15. Sorrentino E, Diedler J, Kasprowicz M, et al (2012) Critical Thresholds for Cerebrovascular Reactivity After Traumatic Brain Injury. Neurocrit Care 16:258-266. https: / / doi.org / 10.1007 / sl2028-011-9630-8

[0114] 16. Stein KY, Froese L, Gomez A, et al (2022) Association between cerebrovascular reactivity in adult traumatic brain injury and improvement in patient outcome over time: an exploratory analysis. Acta Neurochir (Wien) 164:3107-3118. https: / / doi.org / 10.1007 / s00701-022-05366-9

[0115] 17. Zeiler FA, Ercole A, Beqiri E, et al (2019) Cerebrovascular reactivity is not associated with therapeutic intensity in adult traumatic brain injury: a CENTER-TBI analysis. Acta Neurochir 161 : 1955-1964. https: / / doi.org / 10.1007 / s00701-019-03980-8

[0116] 18. Zeiler FA, Ercole A, Beqiri E, et al (2019) Association between Cerebrovascular Reactivity Monitoring and Mortality Is Preserved When Adjusting for Baseline Admission Characteristics in Adult Traumatic Brain Injury: A CENTER-TBI Study. Journal of

[0117] Neurotrauma, https: / / doi.org / 10.1089 / neu.2019.6808

[0118] 19. Gomez A, Froese L, Sainbhi AS, et al (2021) Transcranial Doppler Based Cerebrovascular Reactivity Indices in Adult Traumatic Brain Injury: A Scoping Review of Associations With Patient Oriented Outcomes. Frontiers in Pharmacology 12:1720. https: / / doi.org / 10.3389 / fphar.2021.690921

[0119] 20. Hori D, Nomura Y, Ono M, et al (2017) Optimal blood pressure during cardiopulmonary bypass defined by cerebral autoregulation monitoring. J Thorac Cardiovasc Surg 154:1590-1598. e2. https: / / doi.org / 10.1016 / jjtcvs.2017.04.091

[0120] 21. Joshi B, Brady K, Lee J, et al (2010) Impaired Autoregulation of Cerebral Blood Flow During Rewarming from Hypothermic Cardiopulmonary Bypass and Its Potential Association with Stroke. Anesth Analg 110:321-328. https : / / doi . org / 10.1213 / ANE.0b013 e3181 c6fd 12

[0121] 22. Zeiler FA, Ercole A, Cabeleira M, et al (2019) Comparison of Performance of Different Optimal Cerebral Perfusion Pressure Parameters for Outcome Prediction in Adult Traumatic Brain Injury: A Collaborative European NeuroTrauma Effectiveness Research in Traumatic Brain Injury (CENTER-TBI) Study. J Neurotrauma 36: 1505-1517. https: / / doi.org / 10.1089 / neu.2018.6182

[0122] 23. Aries MJ, Czosnyka M, Budohoski K, et al (2012) Continuous determination of optimal cerebral perfusion pressure in traumatic brain injury*. Critical Care Medicine 40:2456- 2463. https: / / doi.org / 10.1097 / CCM.0b013e3182514eb6

[0123] 24. Malhotra AK, Schweitzer JB, Fox JL, et al (2003) Cerebral perfusion pressure directed therapy following traumatic brain injury and hypotension in swine. Journal of Neurotrauma 20:827-839

[0124] 25. Howells T, Johnson U, McKelvey T, Enblad P (2015) An optimal frequency range for assessing the pressure reactivity index in patients with traumatic brain injury. J Clin Monit Comput 29:97-105. https: / / doi.org / 10.1007 / sl0877-014-9573-7 26. Froese L, Gomez A, Sainbhi AS, et al (2022) Continuous Determination of the Optimal Bispectral Index Value Based on Cerebrovascular Reactivity in Moderate / Severe Traumatic Brain Injury: A Retrospective Observational Cohort Study of a Novel Individualized Sedation Target. Crit Care Explor 4:e0656. https: / / doi.org / 10.1097 / CCE.0000000000000656

[0125] 27. Froese L, Gomez A, Sainbhi AS, et al (2022) Optimal bispectral index level of sedation and cerebral oximetry in traumatic brain injury: a non-invasive individualized approach in critical care? Intensive Care Medicine Experimental 10:33. https: / / doi.org / 10.1186 / s40635-022-00460-9

[0126] 28. Nagaraj SB, McClain LM, Boyle EJ, et al (2018) Electroencephalogram based Detection of Deep Sedation in ICU Patients Using Atomic Decomposition. IEEE Trans Biomed Eng 65:2684-2691. https : / / doi . org / 10.1109 / TBME.2018.2813265

[0127] 29. Rasulo FA, Hopkins P, Lobo FA, et al (2023) Processed Electroencephalogram-Based Monitoring to Guide Sedation in Critically Ill Adult Patients: Recommendations from an International Expert Panel-Based Consensus. Neurocrit Care 38:296-311. https: / / doi.org / 10.1007 / sl2028-022-01565-5

[0128] 30. Mathur S, Patel J, Goldstein S, Jain A (2020) Bispectral Index. In: StatPearls. StatPearls Publishing, Treasure Island (FL)

[0129] 31. Sleigh JW, Andrzejowski J, Steyn-Ross A, Steyn-Ross M (1999) The Bispectral Index: A Measure of Depth of Sleep? Anesthesia & Analgesia 88:659-661. https: / / doi.org / 10.1213 / 00000539-199903000-00035

[0130] 32. Tas J, Beqiri E, van Kaam RC, et al (2021) Targeting Autoregulation-Guided Cerebral Perfusion Pressure after Traumatic Brain Injury (COGiTATE): A Feasibility Randomized Controlled Clinical Trial. Journal of Neurotrauma 38:2790-2800. https: / / doi.org / 10.1089 / neu.2021.0197

[0131] 33. Lassen NA (1964) AUTOREGULATION OF CEREBRAL BLOOD FLOW. Circ Res 15:SUPPL:201-204 34. Pan J, Tompkins WJ (1985) A Real-Time QRS Detection Algorithm. IEEE Transactions on Biomedical Engineering BME-32:230-236. https: / / doi.org / 10.1109 / TBME.1985.325532

[0132] 35. Czosnyka M, Smielewski P, Kirkpatrick P, et al (1997) Continuous Assessment of the Cerebral Vasomotor Reactivity in Head Injury. Neurosurgery 41 : 11-19. https : / / doi. org / 10.1097 / 00006123-199707000-00005

[0133] 36. Zeiler FA, Lee JK, Smielewski P, et al (2018) Validation of Intracranial Pressure-Derived Cerebrovascular Reactivity Indices against the Lower Limit of Autoregulation, Part II: Experimental Model of Arterial Hypotension. J Neurotrauma 35:2812-2819. https: / / doi.org / 10.1089 / neu.2017.5604

[0134] 37. Fischer GW, Torrillo TM, Weiner MM, Rosenblatt MA (2009) The Use of Cerebral Oximetry as a Monitor of the Adequacy of Cerebral Perfusion in a Patient Undergoing Shoulder Surgery in the Beach Chair Position. Pain Practice 9:304-307

[0135] 38. Brady KM, Mytar JO, Kibler KK, et al (2010) Noninvasive Autoregulation Monitoring with and without Intracranial Pressure in the Naive Piglet Brain. Anesth Analg 111:191— 195. https : / / doi .org / 10.1213 / ANE.0b013 e3181 e054ba

[0136] 39. Brady Ken M., Lee Jennifer K., Kibler Kathleen K., et al (2008) Continuous Measurement of Autoregulation by Spontaneous Fluctuations in Cerebral Perfusion Pressure. Stroke 39:2531-2537. https: / / doi.org / 10.1161 / STROKEAHA.108.514877

[0137] 40. Zweifel C, Castellani G, Czosnyka M, et al (2010) Noninvasive Monitoring of Cerebrovascular Reactivity with Near Infrared Spectroscopy in Head-Injured Patients. Journal of Neurotrauma 27: 1951-1958. https: / / doi.org / 10.1089 / neu.2010.1388

[0138] 41. Liu X, Maurits NM, Aries MJH, et al (2017) Monitoring of Optimal Cerebral Perfusion Pressure in Traumatic Brain Injured Patients Using a Multi -Window Weighting Algorithm. J Neurotrauma 34:3081-3088. https: / / doi.org / 10.1089 / neu.2017.5003

[0139] 42. Donnelly J, Czosnyka M, Adams H, et al (2017) Individualizing Thresholds of Cerebral Perfusion Pressure Using Estimated Limits of Autoregulation. Crit Care Med 45:1464- 1471. https: / / doi.org / 10.1097 / CCM.0000000000002575 43. Beqiri E, Smielewski P, Robba C, et al (2019) Feasibility of individualised severe traumatic brain injury management using an automated assessment of optimal cerebral perfusion pressure: the COGiTATE phase II study protocol. BMJ Open 9:. https : / / doi . org / 10.1136 / bmj open-2019-030727

[0140] 44. Zeiler FA, Donnelly J, Menon DK, et al (2018) A Description of a New Continuous Physiological Index in Traumatic Brain Injury Using the Correlation between Pulse Amplitude of Intracranial Pressure and Cerebral Perfusion Pressure. Journal of Neurotrauma 35:963-974. https: / / doi.org / 10.1089 / neu.2017.5241

[0141] 45. Zeiler FA, Donnelly J, CalvielloL, et al (2018) Validation of Pressure Reactivity and Pulse

[0142] Amplitude Indices against the Lower Limit of Autoregulation, Part I: Experimental Intracranial Hypertension. J Neurotrauma 35:2803-2811. https: / / doi.org / 10.1089 / neu.2017.5603

[0143] 46. Batson C, Stein KY, Gomez A, et al (2022) Intracranial Pressure-Derived Cerebrovascular Reactivity Indices, Chronological Age, and Biological Sex in Traumatic Brain Injury: A Scoping Review. Neurotrauma Rep 3:44-56. https: / / doi.org / 10.1089 / neur.2021.0054

[0144] 47. Liu X, Donnelly J, Czosnyka M, et al (2017) Cerebrovascular pressure reactivity monitoring using wavelet analysis in traumatic brain injury patients: A retrospective study. PLoS Med 14:el002348. https: / / doi.org / 10.1371 / journal.pmed.1002348

Claims

The invention clamed is:

1. A computer-implemented method of deriving a personalized optimal physiologic target through automated real-time analysis of an electroencephalogram (EEG) entropy and / or variability metric and a cerebrovascular reactivity metric, said method comprising:(a) acquiring, in real-time, and on an ongoing basis, raw input data comprising: real-time EEG entropy and / or variability metric values derived from real-time computer-automated analysis of real-time EEG measurements of a patient; one or more other ongoing real-time physiological measurements of the patient from which real-time cerebrovascular reactivity metric values are derivable through other-real time computer-automated analysis;(b) on an ongoing basis, subjecting said raw input data to a real-time computer- automated artifact detection and removal process, thereby deriving cleaned input data void of error-inducing artifacts removed from the raw input data;(c) on an ongoing basis, performing said other-real time computer-automated analysis of said one or more other ongoing real-time physiological measurements and deriving said real-time cerebrovascular reactivity metric values therefrom;(d) on an ongoing basis, calculating a first summary metric of a sample set of said real-time EEG entropy and / or variability metric values collected in a given window of time and a second summary metric of a sample set of said real-time cerebrovascular reactivity metric values also collected in said given window of time, and saving said first and second summary metrics for said given window of time as a data point;(d) curating a plurality of said data points over longer windows of time, and for each of said longer windows of time, grouping said data points into bins, grouped by values of the first summary metric across said longer window of time, and saving said grouped data points as a respective data set;(e) for each of said respective data sets, calculating a fitted curve to the grouped data points of said data set, and identifying a lowest point on said fitted curve characterized bya lowest cerebrovascular reactivity metric value, and identifying an EEG entropy and / or variability metric value of said lowest point as a respective approximation of said personalized optimal physiologic target;(f) on an ongoing basis, outputting an updated value of said personalized optimal physiologic target that has been derived based, at least in part, on at least one respective approximation from at least one lapsed one of said longer windows of time.

2. The method of claim 1 wherein said EEG entropy metric is an EEG bispectral index (BIS) and said real-time EEG entropy metric values are BIS values received from a BIS monitor that processes the EEG measurements of the patient.

3. The method of claim 1 or 2 said real-time cerebrovascular reactivity metric is an intracranial pressure (ICP) derived cerebrovascular reactivity metric.

4. The method of claim 3 wherein said real-time cerebrovascular reactivity metric is selected from a group consisting of a pressure reactivity index (PRx), a pulse amplitude index (PAx), and a correlation between pulse amplitude of ICP and cerebral perfusion pressure (RAC).

5. The method of claim 1 or 2 wherein said real-time cerebrovascular reactivity metric is a near infrared spectroscopy (NIRS) derived cerebrovascular reactivity metric.

6. The method of claim 5 wherein said NIRS derived cerebrovascular reactivity metric is selected from a group consisting of a cerebral oxygen index, an oxyhemoglobin (HbO) index (HbOx), a deoxyhemoglobin (Hb) index (Hbx), a tissue oxygen saturation (TOI) index (TOx), and a total hemoglobin (THI) index (THx).

7. The method of any preceding wherein said other one or more physiological measurements comprises at least two different physiological measurements and other-real time computer-automated analysis comprises a Pearson correlation between respective series of consecutive mean values for each of said at least two different physiological measurements.

8. The method of any preceding claim comprising receiving said other one ormore physiological measurements from one or more non-invasive patient monitors.

9. The method of any preceding claim wherein (b) comprises excluding, from the cleaned data, real-time physiological measurements falling outside predetermined thresholds that denote opposite ends of a reasonable measurement range.

10. The method of any preceding claim wherein (b) comprises performing detection of pulse wave forms in at least one of said one or more other ongoing real-time physiological measurements, and excluding from said cleaned input data any measurement signals that both exceed a predetermined length and are void of any said pulse wave forms.

11. The method of any preceding claim wherein (b) comprises monitoring electromyography (EMG) signals for instances of extreme magnitude, and for any EEG entropy and / or variability metric values coinciding with such instances, excluding said any EEG entropy and / or variability values from said cleaned input data.

12. The method of any preceding claim wherein (e) further comprises calculating a curve- strength factor based on difference of magnitude between a lowest cerebrovascular reactivity metric value among the grouped data points of the data set, and a higher one of two cerebrovascular reactivity metric values each corresponding to a respective end of the fitted curve.

13. The method of claim 12 comprising using said curve- strength factor for at least one of the following:(i) weighted determination of the updated value in step (e) based on a plurality of respective approximations from a lapsed plurality of said longer windows of time, which are weighted based at least partly on said curve-strength factor; and / or(ii) output of said curve strength factor, or other readable curve strength indicator derived therefrom, with the updated value in step (f) to give added context to said updated value.

14. The method of any preceding claim wherein (e) further comprises calculating a lowest-point location factor characterizing a locality of the lowest point along the fitted curve relative to a quantity of bins into which the data points were grouped.

15. The method of claim 14 comprising using said lowest-point location factor for at least one of the following:(i) weighted determination of the updated value in step (e) based on a plurality of respective approximations from a lapsed plurality of said longer windows of time, which are weighted based at least partly on said lowest-point location factor; and / or(ii) output of said lowest-point location factor, or other readable locationpertinent indicator derived therefrom, with the updated value in step (f) to give added context to said updated value.

16. The method of any preceding claim wherein step (f) comprises displaying said updated value, or an equivalent readable indicator of said personalized targetable optimal level of sedation, in a local environment containing said patient and / or one or more medical practitioners seeing to said patient.

17. The method of claim 16 wherein said local environment is an intensive care unit.

18. The method of claim 16 wherein said local environment is an operating room.

19. The method of any preceding claim further comprising (g) sending control signals to an intravenous sedation titration device to control titration of sedation dosing thereby within a closed loop targeting the updated value of said personalized targetable optimal level of sedation.

20. The method of any preceding claim wherein said summary metrics are median values of said sample sets.

21. The method of any preceding claim wherein said one or more other ongoing real-time physiological measurements comprise at least two real-time physiological measurements.

22. The method of any preceding claim wherein said optimal physiologic target is a targetable optimal level of sedation, and the EEG measurements of the patient serve as an indicator of objective sedation depth of said patient.

23. An electronic device for deriving a personalized optimal physiologic target through automated real-time analysis of an electroencephalogram (EEG) entropy metric and a cerebrovascular reactivity metric, said device comprising one or more processors and non- transitory computer readable memory coupled thereto, in which there are stored statements and instructions that are executable by said one or more processors and are configured to, when executed, perform at least steps (a) through (f) of any preceding claim.

24. The device of claim 23 further comprising an electronic display, and the statements and instructions are further configured to display the updated value of said personalized optimal physiologic target, or an equivalent readable indicator of said personalized optimal physiologic target, on said electronic display.

25. The device of claim 23 or 24 wherein said personalized optimal physiologic target is a targetable optimal level of sedation.

26. The device of claim any one of claims 23 to 25 in combination with an intravenous titration device connected thereto, wherein the statements and instructions are configured to send control signals to the intravenous titration device to control titration of dosing thereby in a closed feedback loop targeting the updated value of said personalized optimal physiologic target.

27. The device of claim 26 wherein said intravenous titration device is a sedation titration device.

28. Non-transitory computer readable memory having stored therein statements and instructions executable by one or more processors to perform at least steps (a) through (f) of any one of claims 1 to 22.

Images