Diversity encouraged learning of unsupervised LSTM ensemble for neural activity video prediction

TL;DR

This paper introduces an ensemble of LSTM auto-encoders trained with an ensemble-awareness loss to improve neural activity video prediction, enabling better modeling of diverse neural patterns for potential real-time brain stimulation.

Contribution

It proposes a novel ensemble training method for LSTM models that automatically specialize in different neural activity patterns without explicit clustering.

Findings

Achieved significant accuracy in neural activity prediction.

Ensemble models effectively specialize in different neural activity clusters.

End-to-end training of the ensemble improves prediction performance.

Abstract

Being able to predict the neural signal in the near future from the current and previous observations has the potential to enable real-time responsive brain stimulation to suppress seizures. We have investigated how to use an auto-encoder model consisting of LSTM cells for such prediction. Recog- nizing that there exist multiple activity pattern clusters, we have further explored to train an ensemble of LSTM mod- els so that each model can specialize in modeling certain neural activities, without explicitly clustering the training data. We train the ensemble using an ensemble-awareness loss, which jointly solves the model assignment problem and the error minimization problem. During training, for each training sequence, only the model that has the lowest recon- struction and prediction error is updated. Intrinsically such a loss function enables each LTSM model to be adapted to a subset of the training sequences that share similar dynamic behavior. We demonstrate this can be trained in an end- to-end manner and achieve significant accuracy in neural activity prediction.