Towards an ontology of cognitive processes and their neural substrates: A structural equation modeling approach

Abstract

A key challenge in the field of cognitive neuroscience is to identify discriminable cognitive functions, and then map these functions to brain activity. In the current study, we set out to explore the relationships between performance arising from different cognitive tasks thought to tap different domains of cognition, and then to test whether these distinct latent cognitive abilities also are subserved by corresponding "latent" brain substrates. To this end, we tested a large sample of adults under the age of 40 on twelve cognitive tasks as they underwent fMRI scanning. Exploratory factor analysis revealed 4-factor model, dissociating tasks into processes corresponding to episodic memory retrieval, reasoning, speed of processing and vocabulary. An analysis of the topographic covariance patterns of the BOLD-response acquired during each task similarity also converged on four neural networks that corresponded to the 4 latent factors. These results suggest that distinct ontologies of cognition are subserved by corresponding distinct neural networks. OPEN ACCESS Citation: Eich T, Parker D, Gazes Y, Razlighi Q, Habeck C, Stern Y (2020) Towards an ontology of cognitive processes and their neural substrates: A structural equation modeling approach. PLoS ONE 15(2): e0228167. https://doi.org/10.interest really do tap that process, and therefore that brain areas that are commonly activated across these different tasks inferentially index that process.Recently, several groups have used different data-driven approaches to uncover the neural networks underlying different cognitive abilities. Yeo and colleagues [7], for example, leveraged data from the BrainMap database [8], which includes over 10,000 imaging experiments, to create a "latent cognitive structure and its topography" ([7], p. 3665-6). They explored functional specialization and flexibility across 83 different cognitive tasks and reported meaningful functional specialization across tasks (e.g., tasks which in the extant literature are commonly thought to tap inhibitory processes, including the Simon, Posner, Stroop, Flanker, N-Back and Task Switching, all recruited a common neural network), but also reported that individual networks can demonstrate remarkable flexibility (e.g., specific cortical regions participated in multiple cognitive components and divergent cognitive tasks). Bertolero and colleagues [9][10][11] extended this work in several important ways using graph theory network based approaches, first showing with resting state functional MRI data that distinct modules perform distinct cognitive functions, but that, as the number of cognitive functions within a given experimental task increases, so too does activity in connector nodes that link modules [9]. Bertolero and colleagues further showed that hub connectivity accurately predicts performance across 4 different cognitive tasks, and found that those individual who had hubs with higher participation coefficients (e.g., hubs with greater diversity in modular connections) as opposed to higher hub stre...

Reduced cortical thickness in World Trade Center responders with cognitive impairment

“…The entorhinal cortex is concerning because it is commonly influenced by most dementing conditions, and thus the lack of results in the entorhinal cortex in particular is unusual. One interpretation of this finding may be that these regions are commonly reduced across the WTC responder population, but may not be sufficient to cause cognitive symptoms on their own 62 . Results comparing CTX in cognitively unimpaired WTC responders with published normative data from 2,799 cognitively normal adults 27 identified substantial reductions in the entorhinal cortex and across the temporal lobe despite revealing thicker cortices on average across many other regions; results may suggest that cognitively unimpaired WTC responders also have some abnormalities in this region of the brain.…”

Intrinsic Connectivity Changes Mediate the Beneficial Effect of Cardiovascular Exercise on Sustained Visual Attention

“…Cognitive functions can be viewed as the behavioral outcome of collaborative processing of sensory information by distributed but interconnected neural systems ( Mesulam 1998 ). Based on psychometric theory and supported by neuroscientific evidence of specialized neural networks ( Eich et al. 2020 ), cognitive functions can be subdivided into domains such as processing speed, reasoning, and memory, to name but a few ( McGrew 2009 ).…”

Cognitive reserve protects ALS‐typical cognitive domains: A longitudinal study.

“…However, future research needs to determine conclusive evidence for executive, language, social cognition and memory functions across the ALS-FTSD. Given that the architecture of cognition remains a matter of debate [52][53][54] and that atrophy is widespread, 18 future studies should direct their attention towards cognitive networks to investigate the interrelationships between cognitive functions in ALS-FTSD. Establishing cognitive networks may reveal structures underlying these vulnerabilities.…”

“…This absence is likely because neuroscience overall still struggles to map cognitive functions onto brain areas 53 and to establish the architecture of cognitive functions overall. Factor analyses have suggested that memory and reasoning could be one domain 54 and that verbal fluency might be a language function rather than an executive one as it relates to both domains 52 . This necessitates future network analyses, to see if ALS‐FTSD patients and healthy persons differ in their cognitive architecture.…”