Abstract:Objective.Implanted devices providing real-time neural activity classification and control are increasingly used to treat neurological disorders, such as epilepsy and Parkinson's disease. Classification performance is critical to identifying brain states appropriate for the therapeutic action (e.g. neural stimulation). However, advanced algorithms that have shown promise in offline studies, in particular deep learning (DL) methods, have not been deployed on resource-restrained neural implants. Here, we designed and optimized three DL models or edge deployment and evaluated their inference performance in a case study of seizure detection.Approach.A deep neural network (DNN), a convolutional neural network (CNN), and a long short-term memory (LSTM) network were designed and trained with TensorFlow to classify ictal, preictal, and interictal phases from the CHB-MIT scalp EEG database. A sliding window based weighted majority voting algorithm was developed to detect seizure events based on each DL model's classification results. After iterative model compression and coefficient quantization, the algorithms were deployed on a general-purpose, off-the-shelf microcontroller for real-time testing. Inference sensitivity, false positive rate (FPR), execution time, memory size, and power consumption were quantified.Main results.For seizure event detection, the sensitivity and FPR for the DNN, CNN, and LSTM models were 87.36%/0.169 h-1, 96.70%/0.102 h-1, and 97.61%/0.071 h-1, respectively. Predicting seizures for early warnings was also feasible. The LSTM model achieved the best overall performance at the expense of the highest power. The DNN model achieved the shortest execution time. The CNN model showed advantages in balanced performance and power with minimum memory requirement. The implemented model compression and quantization achieved a significant saving of power and memory with an accuracy degradation of less than 0.5%.Significance.Inference with embedded DL models achieved performance comparable to many prior implementations that had no time or computational resource limitations. Generic microcontrollers can provide the required memory and computational resources, while model designs can be migrated to application-specific integrated circuits for further optimization and power saving. The results suggest that edge DL inference is a feasible option for future neural implants to improve classification performance and therapeutic outcomes.

ResOT: Resource-Efficient Oblique Trees for Neural Signal Classification

BrainForest: Neuromorphic Multiplier-Less Bit-Serial Weight-Memory-Optimized 1024-Tree Brain-State Classification Processor

NeuralTree: A 256-Channel 0.227-μJ/Class Versatile Neural Activity Classification and Closed-Loop Neuromodulation SoC

NeuralTree: A 256-Channel 0.227-$μ$J/Class Versatile Neural Activity Classification and Closed-Loop Neuromodulation SoC

Energy-Efficient Seizure Detection Suitable for low-power Applications

A Resource-Efficient and Data-Restricted Training Method Towards Neurological Symptoms Prediction

Edge Deep Learning for Neural Implants

Edge deep learning for neural implants: a case study of seizure detection and prediction

Closed-Loop Neural Prostheses With On-Chip Intelligence: A Review and a Low-Latency Machine Learning Model for Brain State Detection

NeuroCARE: A Generic Neuromorphic Edge Computing Framework for Healthcare Applications.

Neuromorphic Signal Classification Using Organic Electrochemical Transistor Array and Spiking Neural Simulations

Less is More: Reducing Overfitting in Deep Learning for EEG Classification.

A resource-efficient ECG diagnosis model for mobile health devices

Achieving Pareto Optimality using Efficient Parameter Reduction for DNNs in Resource-Constrained Edge Environment

Biological plausible algorithm for seizure detection: Toward AI-enabled electroceuticals at the edge

Energy-efficient dynamic sensor time series classification for edge health devices

An Accurate EEGNet-based Motor-Imagery Brain-Computer Interface for Low-Power Edge Computing

An Energy-Efficient ECG Processor with Ultra-Low-Parameter Multi-Stage Neural Network and Optimized Power-of-Two Quantization

A Comparison of Energy-Efficient Seizure Detectors for Implantable Neurostimulation Devices

Scalable Machine Learning Architecture for Neonatal Seizure Detection on Ultra-Edge Devices

A Compact Multi-Branch 1D Convolutional Neural Network for EEG-based Motor Imagery Classification