Assessing spatial orientation in the human hippocampus during virtual navigation

Abstract

Neural decoding uses electrophysiological signals recorded from the brain to find identifications between these signals and behavioral variables in the outside world. These decoding predictions can be used to better understand how areas of the brain relate to the observable states. In essence, neural decoding is a regression (or classification) problem relating neural signals to particular variables [1]. In this project, we attempt to identify human hippocampal local field potential (LFP) with spatial orientation during virtual navigation, guided by some important results from the literature such as the apparent relation between LFP and information about speed and location in rats [2], the identification between entorhinal cortex and information about head direction, self-location, and speed in rats [3], [4], the relation between thalamus and subiculum with 3D head direction [5] or the association of hippocampal theta oscillations to mnemonic processes [6]. To address this proposal, we use an adaptation of the convolutional neural network proposed by Frey et al. [7] that can be found here. Our results suggest that spatial orientation is encoded in human hippocampal LFP during virtual navigation.