Dataset of human intracranial recordings during famous landmark identification

Abstract

For most people, recalling information about familiar items in a visual scene is an effortless task, but it is one that depends on coordinated interactions of multiple, distributed neural components. We leveraged the high spatiotemporal resolution of direct intracranial recordings to better delineate the network dynamics underpinning visual scene recognition. We present a dataset of recordings from a large cohort of humans while they identified images of famous landmarks (50 individuals, 52 recording sessions, 6,775 electrodes, 6,541 trials). This dataset contains local field potential recordings derived from subdural and penetrating electrodes covering broad areas of cortex across both hemispheres. We provide this pre-processed data with behavioural metrics (correct/incorrect, response times) and electrode localisation in a population-normalised cortical surface space. This rich dataset will allow further investigation into the spatiotemporal progression of multiple neural processes underlying visual processing, scene recognition and cued memory recall.

Fig. 1

Experimental Design and Electrode Coverage. (a) Schematic representation of the landmark identification task. Participants were shown either coherent images of famous landmarks or spatially scrambled versions of the images. (b) Individual electrode locations (6,775 electrodes) and (c) representative coverage map (52 implantations).

Fig. 2

Behavioural Analysis. (a) Number of coherent images presented per participant and the distribution of number of successful naming trials. (b) Response time (RT) distributions for correctly answered responses for coherent and scrambled trials.

Fig. 3

Spatiotemporal Profile of Cortical Activations. Locations (a) and BGA activations (b; mean ± s.e.) of six ROIs based on the HCP parcellation. V1 (Primary visual cortex; 84 electrodes, 15 patients), V4 (Fourth visual area; 56 electrodes, 12 patients), POS1 (Parieto-occipital sulcus area 1; 69 electrodes, 27 patients), 7 m (Area 7 m; 56 electrodes, 20 patients), PHA1 (Parahippocampal Area 1; 27 electrodes, 14 patients), and EC (Entorhinal cortex; 60 electrodes, 25 patients). Vertical dashed lines denote stimulus onset and offset times.