Bionic Brain Region Structural-Functional Division and Event-Driven Synergistic Method

By mapping the functional modules of the brain-like intelligent system to biomimetic brain regions and organizing them according to functional hierarchy, defining standardized interfaces and signal transmission and reception methods, the problems of ambiguous module responsibility boundaries and dispersed global state regulation are solved, and efficient collaboration and unified regulation of biomimetic brain regions are achieved.

CN123114428APending Publication Date: 2026-11-17SHANGHAI YAOSHI TECHNOLOGY CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611471605.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-09-21
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing brain-inspired intelligent systems suffer from vague module responsibilities, inconsistent interfaces, lack of functional basis for startup sequence, and fragmented global state regulation, making it impossible to effectively and systematically map the human brain's anatomical structure into software brain regions.

Method used

The functional modules of the brain-like intelligent system are mapped to multiple biomimetic brain regions. Each region corresponds to a specific biological function in the human brain's anatomical structure. The systems are organized into a hierarchical structure from bottom to top according to predetermined functional levels. Standardized interfaces and signal transmission and reception methods are defined, and event-driven collaboration is achieved through initialization by a startup manager and message router.

Benefits of technology

It achieves the definition of the responsibilities of bionic brain regions, ensures that functions do not overlap or omission, and ensures that the startup sequence and information flow are biologically reasonable. It supports plug-and-play and parallel development by multiple teams and has the ability to uniformly regulate the global state.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123114428A_ABST
    Figure CN123114428A_ABST
Patent Text Reader

Abstract

This invention discloses a biomimetic brain region structure and function division and event-driven collaboration method. Functional modules are mapped to multiple biomimetic brain region units, each corresponding to a prototype brain region in the human brain anatomy, with responsibility boundaries determined by the biological function of the prototype brain region. The brain region units are organized into a hierarchical structure from bottom to top according to functional levels. A standardized interface containing a unified lifecycle management method and signal transmission / reception method is defined for each brain region unit. A startup manager initializes each brain region unit sequentially from the basic layer to the advanced layer according to functional hierarchy and dependencies, and performs state verification. During runtime, event-driven collaboration is achieved by transmitting messages through a message router based on subscription relationships or target brain region identifiers. Brainstem and midbrain nuclei units use modulated signals carrying intensity parameters to achieve global regulation of arousal, emotion, and reward. This invention clarifies responsibility boundaries, unifies interfaces, and ensures biologically rational startup sequences, improving the maintainability and scalability of the brain-like system.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of brain-inspired computing and intelligent robot software architecture, and in particular to a biomimetic brain region structure and function partitioning and event-driven collaboration method, which can be applied to the software architecture design and implementation of humanoid robots, biomimetic intelligent agents and brain-inspired intelligent systems. Background Technology

[0002] With the development of robotics technology, intelligent robot systems are gradually evolving from single-function executors into brain-like intelligent agents with perception, cognition, memory, learning, and decision-making capabilities. How to organize numerous processing modules with diverse functions in the system software architecture has become a key issue in the realization of brain-like intelligent systems.

[0003] The current mainstream technology routes mainly include the following three categories:

[0004] The first category is monolithic deep learning or large-model architectures. These architectures are centered around end-to-end neural networks, large language models, or vision-language-action (VLA) models, coupling perception, cognition, decision-making, and control within a single model. While this type of approach performs well on specific tasks, its functions are not modularized; internally, it is organized by network layers rather than by functional domains, making it impossible to locate errors by function. Adding new capabilities typically requires re-collecting data and retraining the model, resulting in high expansion costs. Furthermore, the architecture cannot be benchmarked against and validated with neuroscience findings, making it difficult to adopt functional organization methods proven effective in brain science research.

[0005] The second type is the layered service architecture. This architecture breaks down the system into service modules such as perception, planning, control, and interaction, which coordinate through remote procedure calls or message topics. Although it achieves a certain degree of modularity, the module division is based on engineering habits or data flow rather than the division of functions in the brain. This leads to blurred boundaries of module responsibilities, with the same function (such as emotion regulation, memory management, and arousal control) being scattered across multiple modules. The interfaces between modules are customized and have different styles, and adding a new module requires adaptation to multiple calling methods. The startup order is only arranged according to engineering dependencies, lacking a functional layering basis, which easily leads to startup idleness caused by higher-level modules waiting for lower-level services.

[0006] The third category is traditional cognitive architectures, such as generative rule architectures and modular cognitive architectures. Although this type of architecture draws on some concepts from cognitive psychology, it only contains a small number of functional modules (usually 5 to 10, such as perception, memory, and decision-making modules). It does not systematically divide the modules according to the anatomical structure of the human brain, and the granularity of the modules is too coarse. It cannot support the engineering implementation of rich brain functions such as emotion, arousal, reward, smell, and motor coordination, nor can it reflect the hierarchical support and regulation relationships between brain regions.

[0007] The above solutions fail to effectively address the following issues: how to systematically map the human brain's anatomical structure into software brain regions, ensuring each region has defined responsibilities and functions that are neither overlapping nor omitted; how to organize brain regions according to functional levels, from the brainstem life support layer to the cognitive cortex, so that the activation sequence and information flow are biologically rational; how to define a unified, standardized message interface for the numerous brain regions, enabling any brain region to be independently developed, tested, replaced, and expanded; and how to coordinate between brain regions in an event-driven manner, and achieve unified regulation of the entire system's state through modulation signals from the brainstem and midbrain nuclei. Summary of the Invention

[0008] This invention provides a biomimetic brain region structure and function division and event-driven coordination method to solve the problems of ambiguous module responsibility boundaries, inconsistent interfaces, lack of functional basis for startup sequence, and scattered global state regulation in existing brain-like intelligent systems.

[0009] It should be noted that the present invention does not protect the concept of "bionic brain region" in a biological sense, but rather the software module division, organization and coordination technology solution for brain-like intelligent systems; the terms "brain region", "brainstem", "thalamus" and "amygdala" used in the specification are only figurative descriptions of software functional modules, and their essence is a software unit with a unified interface, defined responsibility boundaries and life cycle management mechanism.

[0010] To achieve the above objectives, the present invention adopts the following technical solution:

[0011] A biomimetic brain region structural and functional division and event-driven collaborative method, applied to brain-like intelligent systems, includes:

[0012] The functional modules of the brain-like intelligent system are mapped to multiple biomimetic brain region units. Each biomimetic brain region unit corresponds to a prototype brain region in the human brain anatomical structure that has a defined biological function. The responsibility boundary of the biomimetic brain region unit in the system is determined by the biological function of the corresponding prototype brain region.

[0013] The multiple biomimetic brain region units are organized into a hierarchical structure from bottom to top according to a predetermined functional hierarchy. The functional hierarchy includes at least the brainstem basal layer, midbrain layer, thalamus layer, limbic system layer, basal ganglia layer, cerebellum layer, sensory cortex, motor cortex, association cortex, cognitive cortex and neuroplasticity layer.

[0014] A standardized interface is defined for each of the bionic brain regions. The standardized interface includes a unified lifecycle management method and a signal transmission and reception method. The messages transmitted between the bionic brain regions include at least the source brain region identifier, the target brain region identifier, the message type, the message content, the priority, and the timestamp.

[0015] The startup manager initializes each of the bionic brain regions sequentially from the basic layer to the advanced layer according to the functional hierarchy and the dependency relationship between the bionic brain regions, and performs state verification after the initialization of each bionic brain region.

[0016] During operation, the bionic brain region units communicate with each other through a message router based on message type subscription relationships or target brain region identifiers to achieve event-driven collaboration.

[0017] Optionally, the multiple bionic brain region units are organized using a tree-like governance structure, which includes one root node, eight brain region managers, and 20 leaf nodes, totaling 29 core brain region units. The eight brain region managers are the brainstem, thalamus, limbic system, basal ganglia, cerebellum, olfactory system, cerebral cortex, and language area, respectively. As a further optional implementation, the bionic brain region units can be added or removed within a range of 20 to 30 units, depending on the system size.

[0018] Optionally, the functional levels, from bottom to top, are the brainstem basal layer, midbrain layer, thalamus layer, limbic system layer, basal ganglia layer, cerebellum layer, sensory cortex, motor cortex, association cortex, cognitive cortex, neuroplasticity layer, and knowledge cortex and language area, totaling twelve functional levels, wherein the knowledge cortex and language area are at the same functional level; the lower brain regions provide support for the upper brain regions, and the upper brain regions regulate the lower brain regions.

[0019] Optionally, the unified lifecycle management method includes activation method, deactivation method, state update method, signal reception method, state acquisition method, and capability declaration method; the state of the bionic brain region unit includes four states: initialized, active, inactive, and error.

[0020] Optionally, the method further includes: performing interface compliance verification on any of the bionic brain region units, checking whether the bionic brain region unit implements the lifecycle management method, and not allowing bionic brain region units that fail verification to access the system.

[0021] Optionally, the signal types transmitted between the bionic brain regions are named hierarchically and divided into eight categories: perception, cognition, memory, emotion, action, emergency, system, and semantic. The messages are classified into four categories according to their purpose: signal messages, command messages, data messages, and status messages.

[0022] Optionally, the dependencies include: the amygdala unit depends on the thalamic nucleus unit and the hypothalamic unit; the hippocampal unit depends on the amygdala unit and the thalamic nucleus unit; the neuroplasticity unit depends on the hippocampal unit; the occipital lobe unit, the temporal lobe unit, and the parietal lobe unit depend on the thalamic nucleus unit; the prefrontal cortex unit depends on the frontal lobe unit; the association cortex unit depends on the occipital lobe, temporal lobe, parietal lobe, and frontal lobe units; the knowledge cortex unit depends on the temporal lobe, hippocampus, and neuroplasticity unit; and the Broca's area unit depends on the frontal lobe, temporal lobe, and neuroplasticity unit.

[0023] Optionally, the startup manager initializes each of the bionic brain region units sequentially from the basic layer to the advanced layer in eleven startup stages. The eleven startup stages are as follows: brainstem three-layer initialization, thalamus initialization, limbic system three-layer initialization, neuroplasticity layer initialization, basal ganglia initialization, cerebellum initialization, olfactory system initialization, cerebral cortex four-lobes initialization, prefrontal cortex and knowledge cortex initialization, association cortex initialization, and Broca's area initialization.

[0024] Optionally, the startup control parameters of the startup manager include the inter-stage delay duration, the startup timeout duration of a single bionic brain region unit, and the maximum number of retries; bionic brain region units that fail state verification are reinitialized according to the retry strategy, and when the maximum number of retries is exceeded, they are set to an error state and an alarm is issued.

[0025] Optionally, the message router supports three delivery methods: unicast, subscription by message type, and broadcast; when the target brain region identifier of the message is empty, the message router broadcasts the message to all bionic brain regions.

[0026] Optionally, the method further includes: modulating the state of the bionic brain region units of the entire system by brainstem nuclei and midbrain nuclei units using a modulation signal, wherein the modulation signal carries a target brain region identifier and a signal intensity parameter; the modulation signal includes at least one of a reward prediction error signal, an arousal level regulation signal, and an emotional state regulation signal; and the target bionic brain region unit of the modulation signal adjusts its own processing parameters according to the signal intensity parameter after receiving the modulation signal.

[0027] This invention also provides a biomimetic brain region structure and function division and event-driven collaborative system, comprising:

[0028] The brain region segmentation module is used to map the system functional modules into multiple bionic brain region units. Each bionic brain region unit corresponds to a prototype brain region in the human brain anatomy with a defined biological function, and the responsibility boundary of the bionic brain region unit in the system is determined by the biological function of the corresponding prototype brain region.

[0029] The hierarchical organization module is used to organize the multiple bionic brain region units into a hierarchical structure from bottom to top according to a predetermined functional hierarchy. The functional hierarchy includes at least the brainstem basal layer, midbrain layer, thalamus layer, limbic system layer, basal ganglia layer, cerebellum layer, sensory cortex, motor cortex, association cortex, cognitive cortex and neuroplasticity layer.

[0030] The interface specification module is used to define a standardized interface for each of the bionic brain regions. The standardized interface includes a unified lifecycle management method and a signal transmission and reception method. The messages transmitted between the bionic brain regions include at least the source brain region identifier, the target brain region identifier, the message type, the message content, the priority, and the timestamp.

[0031] The startup management module is used to initialize each of the bionic brain regions sequentially from the basic layer to the advanced layer according to the functional hierarchy and the dependency relationship between the bionic brain regions, and to perform state verification after the initialization of each bionic brain region.

[0032] The message routing module is used to deliver messages between the bionic brain regions based on message type subscription relationships or target brain region identifiers during runtime, thereby achieving event-driven collaboration.

[0033] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the above-described biomimetic brain region structure and function division and event-driven coordination method.

[0034] The beneficial effects brought about by the embodiments of the present invention include:

[0035] By mapping functional modules to biomimetic brain regions that correspond one-to-one with the anatomical structure of the human brain, the responsibility boundaries of each unit are determined by the biological functions of the prototype brain region, ensuring no functional overlap or omission. Abnormalities can be quickly located by brain region, identifying the responsible unit. Through a bottom-up organization of twelve functional levels, the activation sequence and information flow are biologically rational, with lower-level support functions ready before higher cognitive functions. Through a unified lifecycle interface and standardized message structure, any brain region unit can be plugged and used immediately, and adding a new brain region does not require modifying the code of other brain regions. Through the modulation signal mechanism of brainstem nuclei and midbrain nuclei, unified regulation of three global states—awakening, emotion, and reward—is achieved. The architecture can be benchmarked against neuroscience findings, supports continuous evolution, and facilitates parallel development by multiple teams according to brain regions. Attached Figure Description

[0036] Figure 1 is a schematic diagram of the overall architecture of the biomimetic brain region structure and function division and event-driven collaborative system provided in an embodiment of the present invention;

[0037] Figure 2 is a schematic diagram of the tree-shaped governance structure of the biomimetic brain region unit provided in an embodiment of the present invention;

[0038] Figure 3 is a schematic diagram of the twelve functional levels provided in an embodiment of the present invention;

[0039] Figure 4 is a schematic diagram of the standardized brain region interface and message structure provided in an embodiment of the present invention;

[0040] Figure 5 is a schematic diagram of the brain region activation sequence and dependency relationship provided in the embodiment of the present invention;

[0041] Figure 6 is a flowchart of message routing and event-driven collaboration provided in an embodiment of the present invention;

[0042] Figure 7 is a schematic diagram of the global state regulation of the modulation signals of the brainstem nuclei and midbrain nuclei provided in an embodiment of the present invention. Detailed Implementation

[0043] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings.

[0044] Example 1

[0045] As shown in Figures 1 and 2, this embodiment provides a biomimetic brain region structure and function division and event-driven collaborative system. This system is deployed in humanoid robot control software and includes a brain region division module, a hierarchical organization module, an interface specification module, a startup management module, and a message routing module.

[0046] The brain region segmentation module maps the system's functional modules to multiple biomimetic brain region units. Each biomimetic brain region unit corresponds to a prototype brain region in the human brain anatomy with a defined biological function. The responsibility boundary of this biomimetic brain region unit in the system is determined by the biological function of the prototype brain region. In this embodiment, the biomimetic brain region units are organized into a tree-like governance structure with 1 root node, 8 brain region managers, and 20 leaf nodes, totaling 29 core brain region units. The 8 brain region managers are the brainstem, thalamus, limbic system, basal ganglia, cerebellum, olfactory system, cerebral cortex, and language area. The 20 leaf nodes and their corresponding anatomical prototypes and responsibilities are as follows:

[0047] (1) Leaf nodes under the brainstem manager: medulla oblongata, which is responsible for the life center, continuously monitors key indicators such as system CPU, memory, temperature, and battery, realizes basic reflex processing, and manages the balance of sympathetic / parasympathetic; pons, which is responsible for resource allocation and coordination, coordinates resource allocation with the midbrain and thalamus, participates in sleep-wake regulation, and is responsible for motor coordination relay; midbrain, which is responsible for sensory integration and rapid response, and is further included according to the anatomical prototype as superior colliculus (visuospatial processing, eye movement and head movement control, multimodal integration), inferior colliculus (frequency analysis, sound source localization, startle reflex), substantia nigra (dopamine model parameter release, reward prediction error calculation, motor automation) and red nucleus (motor coordination, posture control, error correction).

[0048] (2) Leaf nodes under the thalamus manager: thalamus_nuclei (thalamus nucleus), which assumes the role of information relay center and is responsible for the transmission and filtering of sensory information and attention gating.

[0049] (3) Leaf nodes under the limbic system manager: amygdala, responsible for emotion recognition, fear response, threat detection, and emotional memory; hippocampus, responsible for memory encoding (short-term to long-term episodic memory), spatial navigation, and memory consolidation; hypothalamus, responsible for homeostasis maintenance, autonomic regulation, and sleep regulation; neuroplasticity, responsible for synaptic plasticity, structural plasticity, and consolidation of learning outcomes.

[0050] (4) Leaf nodes under the basal ganglia: basal ganglia is responsible for action selection, habit formation, motor inhibition, and emergency cessation. It is further divided into the striatum (receiving cortical input and activating direct / indirect pathways), the globus pallidus (integrating information and inhibiting motor activity) and the subthalamic nucleus (super-direct pathway and rapid inhibitory regulation) according to the anatomical prototype.

[0051] (5) Leaf nodes under the cerebellum: cerebellum, which is responsible for motor coordination, balance control, action learning, error detection and action automation.

[0052] (6) Leaf nodes under the olfactory system manager: olfactory_bulb (olfactory bulb), responsible for odor detection, lateral inhibition, and odor feature signature; olfactory_cortex (olfactory cortex), responsible for primary olfactory recognition, odor value assessment, and olfactory memory storage and retrieval.

[0053] (7) Leaf nodes under the management of the cerebral cortex: occipital lobe, responsible for visual feature extraction, visual information integration, and visual spatial mapping; temporal lobe, responsible for auditory processing, language comprehension, object and face recognition, and memory association; parietal lobe, responsible for spatial perception, attention guidance, and somatosensory processing; frontal lobe, responsible for motor initiation, motor planning, motor programming, and motor monitoring; association cortex, responsible for cross-modal integration, abstract concept formation, and reasoning and decision-making; prefrontal cortex, responsible for working memory, planning, decision generation, inhibitory control, and social cognition; knowledge cortex, responsible for knowledge storage and retrieval.

[0054] (8) Leaf node under the language area manager: broca_area (Broca area), which is responsible for coordinating the language generation and pronunciation process.

[0055] Below the aforementioned 20 leaf nodes, further subdivisions based on anatomical prototypes can be made into nucleus-level units: brainstem nuclei (habenular nucleus, locus coeruleus, raphe nucleus), basal ganglia nuclei (striatum, globus pallidus, subthalamic nucleus), limbic system nuclei (nucleus accumbens, septal nucleus), etc. These nucleus-level units are implemented as internal components of the brain region units within their respective leaf nodes, and do not participate in initialization and message delivery as independent bionic brain region units. The core brain region units in this embodiment are the aforementioned 29. As an optional implementation, bionic brain region units can be added or removed within a range of 20 to 30 units depending on the system scale (e.g., disabling the olfactory bulb and olfactory cortex units in models without olfactory hardware).

[0056] Example 2

[0057] As shown in Figure 3, this embodiment describes the functional hierarchical organization of bionic brain region units.

[0058] The hierarchical organization module organizes the multiple bionic brain region units into twelve functional levels from bottom to top according to a predetermined functional hierarchy:

[0059] Layer 0 is the basic brainstem layer, whose main functions are life maintenance, resource security, and the basis of arousal. The representative brain regions include the medulla oblongata, pons, and the habenula, locus coeruleus, and raphe nuclei, which are realized as internal brainstem nuclei.

[0060] The first layer is the midbrain layer, whose main functions are sensory integration, rapid reflexes, reward modulation, and motor coordination. The representative brain regions include the superior colliculus, inferior colliculus, substantia nigra, and red nucleus.

[0061] The second layer is the thalamus layer, whose main functions are information relay and attention gating. The representative brain region unit is the thalamic nucleus.

[0062] The third layer is the limbic system layer, whose main functions are emotion, memory, and homeostasis. The representative brain regions include the hippocampus, amygdala, hypothalamus, and nucleus accumbens.

[0063] The fourth layer is the basal ganglia layer, whose main functions are action selection and habit formation. The representative brain regions include the striatum, globus pallidus, and subthalamic nucleus.

[0064] The fifth layer is the cerebellum layer, whose main functions are motor coordination and automatic movement. The representative brain region is the cerebellum.

[0065] The sixth layer is the sensory cortex, whose main function is to process primary sensory information. The brain regions represented include the occipital lobe, temporal lobe, parietal lobe, olfactory bulb, and olfactory cortex.

[0066] The 7th layer is the motor cortex, whose main function is motor planning and execution. The representative brain region is the frontal lobe, which is responsible for motor initiation, motor planning, motor programming and motor monitoring, corresponding to the functions of the motor cortex in anatomy.

[0067] The 8th layer is the association cortex, whose main functions are information integration and cross-modal processing. The representative brain region is the association cortex.

[0068] The 9th layer is the cognitive cortex, whose main functions are higher cognition, reasoning and decision-making. The representative brain region is the prefrontal cortex.

[0069] The 10th layer is the neuroplasticity layer, whose main functions are learning, memory consolidation, and synaptic adjustment. The brain region unit represented is the neuroplasticity module.

[0070] Layer 11 is the knowledge cortex and language area, whose main functions are knowledge storage and retrieval and language production. The representative brain regions are the knowledge cortex and Broca's area.

[0071] Information flow between levels includes two pathways: the bottom-up sensory pathway, from external input via the thalamus to the primary sensory cortex, then via the association cortex to the prefrontal cortex; and the rapid reflex pathway, from the brainstem / midbrain directly via the basal ganglia and cerebellum to motor output. Upper brain regions regulate lower brain regions; for example, the prefrontal cortex exerts inhibitory control over the limbic system, and arousal nuclei modulate cortical arousal levels.

[0072] Example 3

[0073] As shown in Figure 4, this embodiment describes the standardized brain region interface and message structure.

[0074] The interface specification module defines a standardized interface for each bionic brain region unit. The unified lifecycle management method includes six methods: the `activate` method, used to activate the brain region unit and return whether activation was successful; the `deactivate` method, used to deactivate the brain region unit; the `update` method, called in each processing cycle, used to handle internal state and logic; the `receive_signal` method, used to receive signals from other brain region units, with input parameters including the name of the source brain region, signal type, signal data, and signal strength, and returning optional response data; the `get_status` method, which returns the status information of the brain region unit; and the `get_capabilities` method, which returns a list of the brain region unit's capability names.

[0075] The state machine of the bionic brain region unit includes four states: initialized, active, inactive, and error. The brain region unit is in the initialized state after construction, enters the active state after successful activation, enters the inactive state after deactivation, and enters the error state when initialization or operation fails.

[0076] The system also provides an interface compliance verification method to check whether any bionic brain region unit implements the above six required methods and whether the methods can be called. Bionic brain region units that fail the verification are not allowed to access the system.

[0077] Above the base class interface, each biomimetic brain region unit defines its domain interface according to its functional prototype. For example, the domain interface of the brainstem unit includes methods for acquiring vital signs, triggering basic reflexes, adjusting autonomic balance, maintaining wakefulness, and emergency response; the domain interface of the habenula unit includes methods for handling reward prediction errors, emotion regulation, and inhibiting the raphe nucleus; the domain interface of the locus coeruleus unit includes methods for releasing norepinephrine model parameters to the target brain region, adjusting wakefulness, handling stress, and regulating attention; the domain interface of the superior colliculus unit includes methods for handling visual input, generating eye movement commands, generating head movement commands, and multimodal integration; and the domain interface of the substantia nigra unit includes methods for releasing dopamine model parameters, calculating reward prediction errors, and acquiring dopamine levels.

[0078] The types of signals transmitted between brain regions are named hierarchically and divided into eight categories with a total of 27 signal types: sensory (sensory.visual, sensory.audio, sensory.tactile, sensory.proprioceptive), cognitive (cognitive.analyze, cognitive.reason, cognitive.learn, cognitive.decide), memory (memory.encode, memory.retrieve, memory.associate, memory.consolidate), emotion (emotion.change, emotion.response, emotion.amplify, emotion.regulate), action (motor.execute, motor.plan, motor.adjust), emergency (emergency.alert, emergency.respond), system (system.status, system.error, system.recover), and semantic (semantic.process, semantic.link, semantic.activate).

[0079] A standardized message structure contains at least six fields: source region identifier (source_region), identifying the brain region from which the message originates; target region identifier (target_region), identifying the target brain region of the message (empty indicates broadcast); message type (message_type), identifying the type of message; message content (content), carrying the message data; priority (priority), identifying the processing priority of the message; and timestamp (timestamp), recording the time the message was generated. Messages are categorized into four types based on their purpose: signal messages, used to transmit short-term activation or inhibition signals (such as reward signals, fear signals); command messages, used to transmit control instructions (such as motor instructions, attention-directed instructions); data messages, used to transmit perceptual information or processing results (such as visual images, auditory spectra); and state messages, used to share the internal states of brain regions (such as emotional states, arousal levels).

[0080] Example 4

[0081] As shown in Figure 5, this embodiment describes the brain region initiation sequences and dependencies.

[0082] The startup management module maintains a dependency table between brain regions. In this embodiment, the dependencies include: no prior dependency on the medulla oblongata; dependency on the medulla oblongata on the pons; dependency on the midbrain on the pons; dependency on the thalamic nuclei on the midbrain; dependency on the hypothalamus on the medulla oblongata and pons; dependency on the amygdala on the thalamic nuclei and hypothalamus; dependency on the hippocampus on the amygdala and thalamic nuclei; dependency on the neuroplasticity module on the hippocampus; dependency on the basal ganglia on the midbrain; no prior dependency on the olfactory bulb; dependency on the olfactory cortex on the olfactory bulb; dependency on the thalamic nuclei on the occipital, temporal, parietal, and frontal lobes; dependency on the prefrontal cortex on the frontal lobe; dependency on the association cortex on the occipital, temporal, parietal, and frontal lobes; dependency on the cognitive cortex on the temporal lobe, hippocampus, and neuroplasticity module; and dependency on the frontal and temporal lobes and neuroplasticity module.

[0083] The startup manager initializes each bionic brain region unit sequentially from the basic layer to the advanced layer in eleven startup stages:

[0084] Phase 1: Initialize the three layers of the brainstem: medulla oblongata, pons, and midbrain;

[0085] Phase 2, Initialization of the thalamus: thalamic nuclei;

[0086] Phase 3: Initialize the three layers of the limbic system: hypothalamus, amygdala, and hippocampus;

[0087] Phase 4, Initialize the neural plasticity layer: the neural plasticity module;

[0088] Phase 5, initialization of the basal ganglia;

[0089] Stage 6, initializing the cerebellum;

[0090] Stage 7, Initialization of the olfactory system: olfactory bulb, olfactory cortex;

[0091] Stage 8: Initialize the four lobes of the cerebral cortex: occipital lobe, temporal lobe, parietal lobe, and frontal lobe;

[0092] Stage 9: Initialize the prefrontal cortex and knowledge cortex;

[0093] Phase 10: Initialize the joint cortex;

[0094] Phase 11, initialize Broca's region (final startup).

[0095] The startup control parameters include: inter-stage delay duration (500 milliseconds in this embodiment); startup timeout duration for a single bionic brain region unit (3000 milliseconds in this embodiment); and maximum number of retries (2 in this embodiment). After each bionic brain region unit is initialized, the startup manager verifies its status to check if the brain region unit has entered an active state; if the verification is successful, it proceeds to the next stage; if the verification fails, it is reinitialized according to the retry strategy; if the maximum number of retries is exceeded, the brain region unit is set to an error state and an alarm is issued.

[0096] The startup sequence follows the developmental sequence of the human brain, which is based on the principle that lower brain regions develop before higher brain regions. This ensures that basic life maintenance and sensory-motor abilities are ready before higher cognitive abilities are ready, avoiding startup idling or failure caused by higher modules waiting for lower-level services.

[0097] Example 5

[0098] As shown in Figure 6, this embodiment describes the message routing and event-driven collaborative process.

[0099] The message routing module provides three delivery methods: unicast, which sends the message to the specified brain region unit according to the target brain region identifier in the message; subscription by message type, where brain regions subscribe to the message types they are interested in, and the message router delivers the matching messages to the subscribers; and broadcast, which delivers the message to all brain regions when the target brain region identifier of the message is empty.

[0100] Taking a sudden noise incident as an example, the collaboration process is as follows:

[0101] Step S601: The inferior colliculus unit receives audio input, completes frequency analysis and sound source localization, determines that a sudden sound triggers the startle reflex, and sends an emergency alarm signal (emergency.alert) to the brainstem.

[0102] Step S602: After receiving the emergency alarm signal, the locus coeruleus unit generates and releases an arousal level adjustment signal carrying a signal strength parameter (0.9 in this example). Each cortical brain region unit increases its own processing gain according to the signal strength in the signal receiving method, thereby improving the arousal level of the entire system.

[0103] Step S603: The temporal lobe unit identifies the sound features and publishes an auditory perception data message (sensory.audio);

[0104] Step S604: The thalamic nucleus unit performs attention gating to route the sound data to the association cortex unit;

[0105] Step S605: The amygdala unit assesses the threat level. If it determines that there is a threat, it issues an emotion response message and amplifies the memory encoding strength in the hippocampus.

[0106] Step S606: The prefrontal cortex unit integrates the above information to make a decision and issues a motion planning command message (motor.plan).

[0107] Step S607: The basal ganglia unit completes the action selection, the cerebellar unit completes the motor coordination, and outputs the motor command to the motor system;

[0108] Step S608: The hippocampal unit performs memory encoding on the current event, and the substantia nigra unit calculates the reward prediction error based on the final execution result and releases a modulation signal.

[0109] All of the above interactions are completed through a message router using standardized messages. Each message carries the source brain region identifier, target brain region identifier, message type, message content, priority, and timestamp, and the interaction process can be fully traced.

[0110] Example 6

[0111] As shown in Figure 7, this embodiment describes the global state regulation mechanism of modulation signals in brainstem nuclei and midbrain nuclei.

[0112] The brainstem nuclei and midbrain nuclei regulate the state of the bionic brain regions of the entire system by modulating signals. The modulating signals carry the target brain region identifier and signal strength parameters ranging from 0.0 to 1.0. After receiving the modulating signals, the target brain regions adjust their processing parameters according to the signal strength parameters.

[0113] The specific adjustment loop includes:

[0114] (1) Habenula-Gap Neural Loop: The habenula unit calculates the reward prediction error based on the difference between the expected value and the actual value. When the actual reward is lower than expected, the activity increases, generating inhibitory input to the gap nucleus unit, reducing the emotional state regulation signal output of the gap nucleus unit, and forming a negative emotion regulation loop.

[0115] (2) Coordination loop between locus coercivity and suture nucleus: The locus coercivity unit and the suture nucleus unit coordinate with each other when regulating arousal, and their activity levels are positively correlated, jointly determining the arousal level of the system;

[0116] (3) Substantia nigra-striatal pathway: The substantia nigra unit calculates the reward prediction error based on the difference between the expected value and the actual value, and sends a reward modulation signal to the striatal unit according to the reward prediction error, driving the direct and indirect pathways of the basal ganglia to achieve action selection and habit formation;

[0117] (4) Downward modulation pathway: Brainstem nucleus units regulate the activity of hypothalamic and limbic system units by sending modulation signals to downstream brain regions, thereby achieving system-wide state regulation.

[0118] Through the above mechanism, the system achieves biomimetic regulation of three global states—awakening, emotion, and reward—in a unified event-driven manner, replacing the decentralized, hard-coded global parameter management in traditional systems.

[0119] Example 7

[0120] This embodiment provides a combined implementation of the present invention and an event-driven multi-level brain region collaborative architecture.

[0121] The biomimetic brain region structure functional division and event-driven collaboration method described in this invention can be implemented in conjunction with other system-level architectures. For example, messages between brain region units can be further encapsulated as system-level events, distributed and scheduled by a unified event bus. The brain region unit responsibility boundaries, functional hierarchies, standardized brain region interfaces, and startup dependency management defined in this invention provide a functional division and organizational basis for event scheduling. The boundary between the two is that the event bus and its distribution and scheduling mechanism belong to a system-level communication architecture; this invention focuses on the division rules, hierarchical organization, interface specifications, and startup management of brain region units.

[0122] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for function division of bionic brain region structure and event-driven cooperation, applied to a brain-like intelligent system, characterized in that, include: The functional modules of the brain-like intelligent system are mapped to multiple biomimetic brain region units. Each biomimetic brain region unit corresponds to a prototype brain region in the human brain anatomical structure that has a defined biological function. The responsibility boundary of the biomimetic brain region unit in the system is determined by the biological function of the corresponding prototype brain region. The multiple biomimetic brain region units are organized into a hierarchical structure from bottom to top according to a predetermined functional hierarchy. The functional hierarchy includes at least the brainstem basal layer, midbrain layer, thalamus layer, limbic system layer, basal ganglia layer, cerebellum layer, sensory cortex, motor cortex, association cortex, cognitive cortex and neuroplasticity layer. A standardized interface is defined for each of the bionic brain regions. The standardized interface includes a unified lifecycle management method and a signal transmission and reception method. The messages transmitted between the bionic brain regions include at least the source brain region identifier, the target brain region identifier, the message type, the message content, the priority, and the timestamp. The startup manager initializes each of the bionic brain regions sequentially from the basic layer to the advanced layer according to the functional hierarchy and the dependency relationship between the bionic brain regions, and performs state verification after the initialization of each bionic brain region. During operation, the bionic brain region units communicate with each other through a message router based on message type subscription relationships or target brain region identifiers to achieve event-driven collaboration.

2. The method according to claim 1, characterized in that, The multiple biomimetic brain region units are organized using a tree-like governance structure, which includes one root node, eight brain region managers, and 20 leaf nodes, totaling 29 core brain region units. The eight brain region managers are the brainstem, thalamus, limbic system, basal ganglia, cerebellum, olfactory system, cerebral cortex, and language area. The functional hierarchy, from bottom to top, consists of twelve levels: the basic brainstem layer, midbrain layer, thalamus layer, limbic system layer, basal ganglia layer, cerebellum layer, sensory cortex, motor cortex, association cortex, cognitive cortex, neuroplasticity layer, and knowledge cortex and language area. The knowledge cortex and language area are at the same functional level. Lower-level brain region units provide support for upper-level brain region units, and upper-level brain region units regulate lower-level brain region units.

3. The method according to claim 1, characterized in that, The unified lifecycle management method includes activation, deactivation, state update, signal reception, state acquisition, and capability declaration methods. The state of the bionic brain region unit includes four states: initialized, active, inactive, and error. The bionic brain region unit is in the initialized state after initialization and enters the active state after successful activation. The method also includes: performing interface compliance verification on any of the bionic brain region units to check whether the bionic brain region unit implements the lifecycle management method. Bionic brain region units that fail verification are not allowed to access the system.

4. The method according to claim 1, characterized in that, The signal types transmitted between the bionic brain regions are named hierarchically and divided into eight categories: perception, cognition, memory, emotion, action, emergency, system, and semantic. The messages are classified into four categories according to their purpose: signal messages, command messages, data messages, and status messages. Signal messages are used to transmit short-term activation or inhibition signals, command messages are used to transmit control instructions, data messages are used to transmit perceptual information or processing results, and status messages are used to share the internal state of the brain regions.

5. The method according to claim 1, characterized in that, The dependencies include: the amygdala unit depends on the thalamic nucleus unit and the hypothalamic unit; the hippocampal unit depends on the amygdala unit and the thalamic nucleus unit; the neuroplasticity unit depends on the hippocampal unit; the occipital lobe unit, the temporal lobe unit, and the parietal lobe unit depend on the thalamic nucleus unit; the prefrontal cortex unit depends on the frontal lobe unit; the association cortex unit depends on the occipital lobe, temporal lobe, parietal lobe, and frontal lobe units; the knowledge cortex unit depends on the temporal lobe, hippocampus, and neuroplasticity unit; and the Broca's area unit depends on the frontal lobe, temporal lobe, and neuroplasticity unit.

6. The method according to claim 1, characterized in that, The startup manager initializes each of the bionic brain region units sequentially from the basic layer to the advanced layer in eleven startup stages. The eleven startup stages are as follows: brainstem three-layer initialization, thalamus initialization, limbic system three-layer initialization, neuroplasticity layer initialization, basal ganglia initialization, cerebellum initialization, olfactory system initialization, cerebral cortex four-lobes initialization, prefrontal cortex and knowledge cortex initialization, association cortex initialization, and Broca's area initialization. The startup control parameters of the startup manager include the inter-stage delay duration, the startup timeout duration of a single bionic brain region unit, and the maximum number of retries. The state verification verifies whether the state of the bionic brain region unit has entered an active state. Bionic brain region units that fail state verification are reinitialized according to the retry strategy. If the maximum number of retries is exceeded, the bionic brain region unit is set to an error state and an alarm is issued.

7. The method according to claim 1, characterized in that, The message router supports three delivery methods: unicast, subscription by message type, and broadcast. When the target brain region identifier of the message is empty, the message router broadcasts the message to all bionic brain regions.

8. The method according to claim 1, characterized in that, Also includes: The system employs modulated signals to regulate the state of all bionic brain regions within the entire system, using brainstem and midbrain nucleus units. These modulated signals carry target brain region identifiers and signal intensity parameters. The modulated signals include at least one of a reward prediction error signal, an arousal level regulation signal, and an emotional state regulation signal. Upon receiving the modulated signals, the target bionic brain region units adjust their processing parameters according to the signal intensity parameters. Specifically, the reward prediction error signal is calculated and generated by the habenular nucleus unit or the substantia nigra unit based on the difference between the expected and actual values; the arousal level regulation signal is generated by the locus coeruleus unit; and the emotional state regulation signal is generated by the raphe nucleus unit.

9. A biomimetic brain region structural and functional division and event-driven collaborative system, characterized in that, include: The brain region segmentation module is used to map the system functional modules into multiple bionic brain region units. Each bionic brain region unit corresponds to a prototype brain region in the human brain anatomy with a defined biological function, and the responsibility boundary of the bionic brain region unit in the system is determined by the biological function of the corresponding prototype brain region. The hierarchical organization module is used to organize the multiple bionic brain region units into a hierarchical structure from bottom to top according to a predetermined functional hierarchy. The functional hierarchy includes at least the brainstem basal layer, midbrain layer, thalamus layer, limbic system layer, basal ganglia layer, cerebellum layer, sensory cortex, motor cortex, association cortex, cognitive cortex and neuroplasticity layer. The interface specification module is used to define a standardized interface for each of the bionic brain regions. The standardized interface includes a unified lifecycle management method and a signal transmission and reception method. The messages transmitted between the bionic brain regions include at least the source brain region identifier, the target brain region identifier, the message type, the message content, the priority, and the timestamp. The startup management module is used to initialize each of the bionic brain regions sequentially from the basic layer to the advanced layer according to the functional hierarchy and the dependency relationship between the bionic brain regions, and to perform state verification after the initialization of each bionic brain region. The message routing module is used to deliver messages between the bionic brain regions based on message type subscription relationships or target brain region identifiers during runtime, thereby achieving event-driven collaboration.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the biomimetic brain region structure and function division and event-driven collaborative method as described in any one of claims 1 to 8.