MTLinteract

The role of neural interaction for context dependent visual processing in the human medial temporal lobe

01.07.2023 bis 30.09.2026

Cognition arises from the coordinated activity of multiple, interacting neural populations at various temporal and spatial scales. The human Medial Temporal Lobe (MTL) comprises anatomically and functionally diverse regions at the interface of perception and memory such as the parahippocampal cortex (PHC), hippocampus (HPC), entorhinal cortex (EC) and amygdala (AM). These areas are heavily reciprocally connected and form several neural loops between neocortex (PHC) and phylogenetically older allocortical (HPC/EC) and subcortical (AM) areas. Each area is functionally specialized: for visuo- spatial processing of complex scenes (PHC), processing of emotion/valence (AM), temporal information (EC) as well as representation of abstract, conceptual knowledge in HPC. However, all areas have also been implicated in context-dependent perceptual processing. As this is thought to rely heavily on top- down influences via feedback projections, MTL regions are well suited for investigating the effect of context-dependent dynamic neuronal interactions. However, surprisingly little is known on how MTL regions interact during context-dependent visual perception.
Here, we will leverage large-scale multi-area recordings of single-unit activity and local field potentials (LFP) in four areas of the human MTL in epilepsy patients during presurgical evaluation while they perform behavioral experiments. We investigate how contextual information influences neural interaction with a specific focus in dissecting feedforward and feedback contributions of each region using advanced data analysis and computational modeling. The first study focuses on how changes in visuo-spatial context affects the interaction between a neocortical region specialized in processing complex visual scenes (PHC) and the hippocampal formation in a scene recognition task. Is context information inherent in PHC activity and transferred to hippocampus in a feedforward way, or relayed back to the PHC via hippocampal feedback loops? The second study addresses how behavioral task context influences feedforward and feedback interactions between MTL regions and whether this is related to their known functional specialization, even though stimulus input remains the same. We will quantify neural interactions using pairwise univariate analyses of spike trains and LFPs to investigate oscillatory synchrony between areas. Moreover, we will employ multivariate measures using dimensionality-reduction methods to discern how feedforward and feedback signalling contributes to contextual modulation at the neuronal population level.
Lastly, using data-constrained deep neural network modelling of multi-region neural activity, we will be able to assess the directed functional interactions between these regions. Taken together, our approach will reveal unique insights into underlying neural interactions as a function of contextual modulation – whether task or stimulus context
- in visual perception and memory at the single neuron level in humans.