synthetic-biological-intelligence - SKILL.md Agent Skill

name: synthetic-biological-intelligence description: "Synthetic Biological Intelligence (SBI) methodology — engineered systems where living Biological Neural Networks (BNNs) are interfaced with hardware/software for task-oriented information processing. Combines organoid technology, MEAs, neuromorphic computing, and ML. Activation: SBI, organoid intelligence, biological neural network, bio-digital interface, wetware computing." version: 1.0.0 author: Research Synthesis license: MIT metadata: hermes: tags: [neuroscience, synthetic-biology, organoid-intelligence, neuromorphic, bio-digital] source_paper: "Synthetic Biological Intelligence: System-Level Abstractions and Adaptive Bio-Digital Interaction (arXiv:2604.27933)" citations: 0 published: "2026-04-30"

Synthetic Biological Intelligence (SBI)

Groundbreaking research paradigm combining organoid technology, Microelectrode Arrays (MEAs), neuromorphic computing, and machine learning to create engineered systems where living Biological Neural Networks (BNNs) are interfaced with hardware and software for task-oriented information processing.

Metadata

Source: arXiv:2604.27933
Authors: Martin Schottlender, Pengjie Zhou, Veronika Volkova
Published: 2026-04-30
Categories: cs.ET (Emerging Technologies)

Core Methodology

What is Synthetic Biological Intelligence?

SBI represents a convergence of multiple fields:

Organoid Technology: 3D cultured brain organoids with functional neural networks
Microelectrode Arrays (MEAs): High-density electrode interfaces for bidirectional communication
Neuromorphic Computing: Hardware that mimics neural architectures
Machine Learning: Algorithms for interpreting and guiding biological computation

System-Level Abstractions

Biological-Digital Interface Architecture

┌─────────────────────────────────────────────────────────┐
│               SBI System Architecture                    │
├─────────────────────────────────────────────────────────┤
│                                                          │
│  ┌─────────────────┐         ┌──────────────────┐       │
│  │  Biological NN   │ ←────→ │    MEA Interface  │       │
│  │  (Organoid/BNN)  │  Bidir │  (Stim/Record)    │       │
│  └─────────────────┘         └────────┬─────────┘       │
│                                      │                  │
│                            ┌─────────▼─────────┐       │
│                            │  Signal Processing │       │
│                            │  (Spike Sorting,   │       │
│                            │   Feature Extract) │       │
│                            └─────────┬─────────┘       │
│                                      │                  │
│                            ┌─────────▼─────────┐       │
│                            │  ML Decoder/Encoder│       │
│                            │  (Classification,  │       │
│                            │   Closed-Loop Ctrl)│       │
│                            └─────────┬─────────┘       │
│                                      │                  │
│                            ┌─────────▼─────────┐       │
│                            │  Task/Application   │       │
│                            │  (Control, Sensing, │       │
│                            │   Computation)      │       │
│                            └─────────────────────┘       │
│                                                          │
└─────────────────────────────────────────────────────────┘

Key Technical Components

1. Biological Neural Network Cultivation

Brain Organoids: Self-organizing 3D neural tissue from stem cells
Functional Maturation: Development of spontaneous and evoked activity
Network Plasticity: Adaptation through stimulation protocols

2. Bidirectional MEA Interface

High-Density Recording: Simultaneous multi-channel neural activity capture
Electrical Stimulation: Precise spatiotemporal pattern delivery
Closed-Loop Control: Real-time feedback between recording and stimulation

3. Signal Processing Pipeline

def process_neural_signals(raw_mea_data):
    """
    Process MEA recordings for SBI systems
    
    Steps:
    1. Spike detection and sorting
    2. Feature extraction (firing rates, burst patterns)
    3. Population activity analysis
    4. State estimation
    """
    # Spike sorting
    spikes = detect_and_sort_spikes(raw_mea_data)
    
    # Extract population features
    firing_rates = compute_firing_rates(spikes, bin_size=50e-3)
    burst_patterns = detect_bursts(spikes)
    
    # Population state
    population_state = compute_population_vector(spikes)
    
    return {
        'spikes': spikes,
        'firing_rates': firing_rates,
        'bursts': burst_patterns,
        'population_state': population_state
    }

4. Machine Learning Integration

Decoding: Map neural activity patterns to intended outputs
Encoding: Convert task inputs into stimulation patterns
Adaptation: Online learning to improve interface performance

Adaptive Bio-Digital Interaction

Closed-Loop Learning Framework

class SBIClosedLoop:
    """
    Adaptive bio-digital interaction system
    
    Continuously adapts stimulation patterns based on
    biological network responses to optimize task performance.
    """
    
    def __init__(self, mea_interface, ml_model, task_env):
        self.mea = mea_interface
        self.ml = ml_model
        self.task = task_env
        self.state = None
        
    def step(self):
        """One step of closed-loop interaction"""
        # 1. Record neural activity
        neural_data = self.mea.record()
        
        # 2. Decode neural state
        self.state = self.ml.decode(neural_data)
        
        # 3. Compute task response
        action = self.task.compute(self.state)
        
        # 4. Encode as stimulation pattern
        stim_pattern = self.ml.encode(action)
        
        # 5. Deliver stimulation
        self.mea.stimulate(stim_pattern)
        
        # 6. Update ML model (online learning)
        self.ml.update(neural_data, action, self.task.reward)
        
        return action

Applications

1. Bio-Hybrid Computing

Novel computational substrate using living neural tissue
Energy-efficient information processing
Adaptive learning capabilities

2. Neuroprosthetics

Advanced brain-machine interfaces
Adaptive prosthetic control
Sensory feedback restoration

3. Drug Screening

Organoid-based pharmacological testing
Real-time neural response monitoring
Personalized medicine applications

4. Fundamental Neuroscience

In vitro models of brain function
Studying learning and memory mechanisms
Disease modeling and intervention

Implementation Challenges

Biological Challenges

Organoid Variability: Batch-to-batch differences in organoid development
Long-term Stability: Maintaining healthy cultures over extended periods
Scalability: Limited size and complexity of current organoids

Technical Challenges

Signal Quality: Noise and artifact rejection in MEA recordings
Latency: Real-time processing requirements for closed-loop systems
Interface Biocompatibility: Long-term electrode-tissue interface stability

Ethical Considerations

Consciousness Questions: At what point might organoid systems develop sentience?
Regulatory Framework: Lack of established guidelines for bio-hybrid systems
Public Perception: Communication about the nature and purpose of SBI

Related Skills

ember-hybrid-snn-llm-architecture
neural-digital-twins-bci
neuromorphic-continual-nuclear-ics
brain-digital-twins-execution-semantics

References

Schottlender, M., Zhou, P., & Volkova, V. (2026). Synthetic Biological Intelligence: System-Level Abstractions and Adaptive Bio-Digital Interaction. arXiv:2604.27933.

Activation Keywords

synthetic biological intelligence, SBI, organoid intelligence, biological neural network, bio-digital interface, wetware computing, brain organoid MEA, neuromorphic biology, bio-hybrid systems, living neural computation