Independent component analysis reveals an early hippocampal recognition system using intracerebral evoked potentials in humans

Abstract

The role of the hippocampal formation in memory recognition has been well studied in animals, with different pathways and structures linked to specific memory processes. In contrast, the hippocampus is commonly analyzed as a unique responsive area in most electrophysiological studies in humans, and the specific activity of its subfields remains unexplored. We combined intracerebral electroencephalogram recordings from epileptic patients with independent component analysis (ICA) during a memory recognition task involving the recognition of old and new images to disentangle the activities of multiple neuronal sources within the hippocampus. We identified two sources with different responses emerging from the hippocampus: a fast one (maximum at [~]250 ms) that could not be directly identified from raw recordings, and a later one, peaking at [~]400 ms. The earliest component was found in 12 out of 15 electrodes, with different amplitudes for old and new items in half of the electrodes. The latter component, identified in 13 out of 15 electrodes, had different delays for each condition, with a faster activation ([~]290 ms after stimulus onset) for recognized items. We hypothesize that both sources represent two steps of hippocampal memory recognition, the faster reflecting the input from other structures and the latter the hippocampal internal processing. Recognized images evoking early activations would facilitate neural computation in the hippocampus, accelerating memory retrieval of complementary information. Overall, our results suggest that hippocampal activity is composed by several sources, including an early system for memory recognition, that can be disentangled with ICA methods. SIGNIFICANCE STATEMENTIn the human memory circuit, the hippocampus is considered as a relatively late structure, associated to the retrieval of elaborated memories. In most electrophysiological studies, it is analyzed as a unique responsive area, and the specific activity of its subfields remains unexplored. In this work, we combined intracerebral recordings with independent component analysis to separate the electrophysiological activity from two different substructures of the hippocampus. We analyzed the responses of both sources in a memory task involving the recognition of old and new images. Our results revealed new hippocampal dynamics associated to different subfields, with memory recognition occurring much faster than previously reported. Importantly, we confirmed the potential of independent component analysis, which can be extended to other brain areas.