Abstract:Objective. Neural decoding is an important tool in neural engineering and neural data analysis. Of various machine learning algorithms adopted for neural decoding, the recently introduced deep learning is promising to excel. Therefore, we sought to apply deep learning to decode movement trajectories from the activity of motor cortical neurons. Approach. In this paper, we assessed the performance of deep learning methods in three different decoding schemes, concurrent, time-delay, and spatiotemporal. In the concurrent decoding scheme where the input to the network is the neural activity coincidental to the movement, deep learning networks including artificial neural network (ANN) and long-short term memory (LSTM) were applied to decode movement and compared with traditional machine learning algorithms. Both ANN and LSTM were further evaluated in the time-delay decoding scheme in which temporal delays are allowed between neural signals and movements. Lastly, in the spatiotemporal decoding scheme, we trained convolutional neural network (CNN) to extract movement information from images representing the spatial arrangement of neurons, their activity, and connectomes (i.e. the relative strengths of connectivity between neurons) and combined CNN and ANN to develop a hybrid spatiotemporal network. To reveal the input features of the CNN in the hybrid network that deep learning discovered for movement decoding, we performed a sensitivity analysis and identified specific regions in the spatial domain. Main results. Deep learning networks (ANN and LSTM) outperformed traditional machine learning algorithms in the concurrent decoding scheme. The results of ANN and LSTM in the time-delay decoding scheme showed that including neural data from time points preceding movement enabled decoders to perform more robustly when the temporal relationship between the neural activity and movement dynamically changes over time. In the spatiotemporal decoding scheme, the hybrid spatiotemporal network containing the concurrent ANN decoder outperformed single-network concurrent decoders. Significance. Taken together, our study demonstrates that deep learning could become a robust and effective method for the neural decoding of behavior.

Continuous Neural Decoding Method Based on General Regression Neural Network.

Neural Decoding Based on Probabilistic Neural Network

Probabilistic neural network based motor cortical decoding method and hardware implementation

Neural Decoding Using a Parallel Sequential Monte Carlo Method on Point Processes with Ensemble Effect.

Temporal Spiking Generative Adversarial Networks for Heading Direction Decoding

Motorsrnn: A Spiking Recurrent Neural Network Inspired by Brain Topology for the Effective and Efficient Decoding of Cortical Spike Trains

Continuous Attractors and Population Decoding Multiple-Peaked Activity

Decoding the Nonstationary Neural Activity in Motor Cortex for Brain Machine Interfaces

Neural Decoding Using Local Field Potential Based on Partial Least Squares Regression.

Robust neural decoding by kernel regression with Siamese representation learning

Decoding Continuous Character-based Language from Non-invasive Brain Recordings

Deep learning for neural decoding in motor cortex

Tracking Functional Changes in Nonstationary Signals with Evolutionary Ensemble Bayesian Model for Robust Neural Decoding

Comparisons Between Linear And Nonlinear Methods For Decoding Motor Cortical Activities Of Monkey

A Sparsity-Adapted Hardware Implementation of SNN for Cortical Spike Trains Decoding

Nonlinear Decoding of Natural Images From Large-Scale Primate Retinal Ganglion Recordings

Generalized neural decoders for transfer learning across participants and recording modalities

GCNs-Net: A Graph Convolutional Neural Network Approach for Decoding Time-resolved EEG Motor Imagery Signals

Continuous Prediction of Mice Lever-Pressing Kinematic Parameters by Background Removed Single-photon Calcium Images

Efficient Neural Decoding Based on Multimodal Training