Deep learning based decoding of local field potential events

How is information processed in the cerebral cortex? To answer this question a lot of effort has been undertaken to create novel and to further develop existing neuroimaging techniques. Thus, a high spatial resolution of fMRI devices was the key to exactly localize cognitive processes. Furthermore, an increase in time-resolution and number of recording channels of electro-physiological setups has opened the door to investigate the exact timing of neural activity. However, in most cases the recorded signal is averaged over many (stimulus) repetitions, which erases the fine-structure of the neural signal. Here, we show that an unsupervised machine learning approach can be used to extract meaningful information from electro-physiological recordings on a single-trial base. We use an auto-encoder network to reduce the dimensions of single local field potential (LFP) events to create interpretable clusters of different neural activity patterns. Strikingly, certain LFP shapes correspond to latency differences in different recording channels. Hence, LFP shapes can be used to determine the direction of information flux in the cerebral cortex. Furthermore, after clustering, we decoded the cluster centroids to reverse-engineer the underlying prototypical LFP event shapes. To evaluate our approach, we applied it to both neural extra-cellular recordings in rodents, and intra-cranial EEG recordings in humans. Finally, we find that single channel LFP event shapes during spontaneous activity sample from the realm of possible stimulus evoked event shapes. A finding which so far has only been demonstrated for multi-channel population coding.