Poster No:
923
Submission Type:
Abstract Submission
Authors:
Sneha Shashidhara1, Moataz Assem2, John Duncan3
Institutions:
1Ashoka University, Sonipat, Haryana, 2MRC Cognition and Brain Sciences Unit, University of Cambridge, Cambridge, Cambridgeshire, 3MRC Cognition and Brain Sciences Unit, Cambridge, Cambridgeshire
First Author:
Co-Author(s):
Moataz Assem
MRC Cognition and Brain Sciences Unit, University of Cambridge
Cambridge, Cambridgeshire
John Duncan
MRC Cognition and Brain Sciences Unit
Cambridge, Cambridgeshire
Introduction:
In the human brain, "multiple-demand" or MD regions are characterised by increased activity associated with many different kinds of cognitive demand (Duncan & Owen, 2000; Duncan et al., 2020), with core components in lateral frontal, dorsomedial frontal and lateral parietal cortex, and multivariate activity patterns that discriminate the contents of many cognitive activities. Using data from 449 participants in the Human Connectome Project (HCP) (Glasser et al., 2016), Assem et al., 2020 found overlapping activity for three types of cognitive demand was strongest in a set of 10 core regions per hemisphere, distributed over the lateral frontal, dorsomedial frontal, insular and lateral parietal cortex MD. Frontal single neuron data show strong selectivity for cognitive operation but weaker selectivity for the stimulus on which this operation is conducted. They also show mixed selectivity; activity depends on the conjunction of multiple task features, such as a particular stimulus object presented at a particular place in a memory list or a particular move planned at a specific position in a sequence (Warden & Miller, 2010; Mushiake et al., 2006; Sigala et al., 2008). Here, we searched for similar properties in fMRI data from core MD regions.
Methods:
Using the advanced fMRI methods of the Human Connectome Project (HCP) and their 360-region cortical parcellation, we conducted a study with 50 human subjects, 37 of whom were included in the analysis. The study consisted of three visual executive tasks intermixed within a scanning run and four scanning runs. Tasks were: n-back working memory (WM), task-switching, and stop-signal, variations of which have previously been shown to recruit the MD network (Fedorenko et al., 2013). The eight blocks of each task consisted of four hard and four easy blocks, two of each using faces and two buildings. We obtained the multivoxel activation pattern for each participant in each of the 360 cortical parcels, and averaged them across cortical networks of core MD and 12 resting-state networks (Ji et al., 2019). To measure dissimilarity between activation patterns we used linear discriminant contrast (LDC) (Nili et al., 2014; Carlin & Kriegeskorte, 2017). We calculated three means for each parcel from the full matrix of LDC distances. STDS was calculated as the mean distance between the face and building blocks in the same task, averaged across tasks. DTSS was calculated as the mean distance between tasks holding stimulus category constant, e.g., n-back face and task switch face averaged across all task pairs and face and building blocks. DTDS was calculated as the mean distance between conditions differing in task and stimulus category, averaged across six such distances. The nonlinearity index was calculated as (STDS+DTSS)-DTDS.
Results:
Stimulus discrimination was strongest in a large region of the early and higher visual cortex. Weak discrimination in core MD regions was accompanied by somewhat stronger discrimination close by in the lateral parietal and frontal cortex, in two bands of regions ventral to the MD core. For DTSS, in contrast, discrimination was strong in the MD core and adjacent regions. For DTDS, as expected, discrimination was widespread, reflecting the union of patterns for STDS and DTSS.
Conclusions:
Core MD had one of the highest non-linearity indices in the brain, albeit a small index. There was scant evidence for mixed selectivity; throughout the brain, discriminations of task and stimulus combined almost linearly. In MD regions, human fMRI data recapitulate some but not all aspects of electrophysiological data from nonhuman primates.
Higher Cognitive Functions:
Executive Function, Cognitive Control and Decision Making 1
Modeling and Analysis Methods:
Multivariate Approaches 2
Keywords:
Cognition
Cortex
Multivariate
1|2Indicates the priority used for review
Provide references using author date format
Assem, M. (2020). A Domain-General Cognitive Core Defined in Multimodally Parcellated Human Cortex. Cerebral Cortex, 30, 4361–4380
Carlin, J. D. (2017). Adjudicating between face-coding models with individual-face fMRI responses. PLOS Computational Biology, 13(7), e1005604.
Duncan, J. (2020). Integrated Intelligence from Distributed Brain Activity. In Trends in Cognitive Sciences (Vol. 24, Issue 10, pp. 838–852)
Duncan, J. (2000). Common regions of the human frontal lobe recruited by diverse cognitive demands. Trends in Neurosciences, 23(10), 475–483.
Fedorenko, E. (2013). Broad domain generality in focal regions of frontal and parietal cortex. Proc. NatL Acad. Sci. USA, 110(41), 16616–16621.
Glasser, M. F. (2016). A multi-modal parcellation of human cerebral cortex. Nature, 536, 171–178.
Ji, J. L. (2019). Mapping the human brain’s cortical-subcortical functional network organization. NeuroImage, 185, 35–57.
Mushiake, H. (2006). Activity in the Lateral Prefrontal Cortex Reflects Multiple Steps of Future Events in Action Plans. Neuron, 50(4), 631–641.
Nili, H. (2014). A Toolbox for Representational Similarity Analysis. PLoS Computational Biology, 10(4), e1003553.
Sigala, N. (2008). Hierarchical coding for sequential task events in the monkey prefrontal cortex. Proc. NatL Acad. Sci. USA, 105(33), 11969–11974.
Warden, M. R. (2010). Task-Dependent Changes in Short-Term Memory in the Prefrontal Cortex. Journal of Neuroscience, 30(47), 15801–15810.