Deep Neural Imputation: A Framework for Recovering Incomplete Brain Recordings

We study the problem of time series imputation in multivariate neural recordings. Compared to standard time series imputation settings, new challenges for imputing neural recordings include the lack of adjacent timestamps for electrodes missing over days, and generalization across days and participants with different electrode configurations. Due to these challenges, the standard practice in neuroscience is to discard electrodes with missing data, even if only a part of the recording is corrupted, significantly reducing already difficult to obtain data. We propose Deep Neural Imputation (DNI), a framework to recover missing electrode recordings by learning across sessions, spatial locations, and participants. We first instantiate DNI with natural linear baselines, then develop encoder-decoder approaches based on masked electrode modeling. We evaluate DNI on 12 distinct multielectrode, naturalistic-behavior human neural datasets, even though canonically in neuroscience, only a few hours of a single participant's recordings would be treated as a distinct dataset to be analyzed separately. Finally, we demonstrate DNI's data imputation ability across many metrics and integrate DNI into an existing neural decoding, data analysis pipeline.