Thalamic regulation of frontal interactions in human cognitive flexibility

Hummos, Ali; Wang, Bin A.; Drammis, Sabrina; Halassa, Michael M.; Pleger, Burkhard · 2022 · Crossref

DOI: 10.1371/journal.pcbi.1010500

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This study investigates the computational role of the mediodorsal thalamus (MD) in regulating interactions within the human frontal cortex, specifically regarding cognitive flexibility. While animal research suggests the MD sustains and switches cortical representations, its precise computational functions and relevance to human decision-making remained unclear. The authors aimed to determine if the MD acts as a functional mediator for frontal cortical areas and to identify the specific mechanisms by which it facilitates task switching and strategy selection. To address this, the researchers combined computational modeling with human functional magnetic resonance imaging (fMRI) data. They extended a neural model of an executive frontal-MD network, originally derived from rodent studies, to simulate a probabilistic inference task performed by human participants. The model featured a recurrent neural network representing the dorsolateral prefrontal cortex (dlPFC) and a layer of MD neurons devoid of local excitatory connections, utilizing biologically plausible Hebbian learning rules. The model was trained to match human behavioral performance on a task requiring subjects to switch between "match" and "non-match" response rules based on probabilistic cue-reward associations. The authors also employed Dynamic Causal Modeling (DCM) on fMRI data to test competing hypotheses about information flow between the dlPFC, orbitofrontal cortex (OFC), and MD. The results demonstrated that the neural model replicated human behavioral dynamics, including flexible switching at block changepoints. Within the model, MD neurons learned to compress dlPFC inputs into abstract representations of the temporal context (dominant strategy). This thalamic compression allowed for efficient partitioning of cortical activity patterns. Perturbation analysis revealed that while steady-state performance remained intact without MD output, the ability to switch strategies rapidly was significantly impaired, indicating the MD’s critical role in transient switching. Furthermore, simulations comparing direct corticocortical pathways versus transthalamic routes (OFC-MD-dlPFC) showed that the thalamic route required fewer neurons and shorter signals to switch dlPFC activity. Human fMRI data supported these findings, showing that sensory inputs entered the circuit via the dlPFC rather than the MD or OFC, confirming the MD’s non-relay role. Additionally, DCM analysis provided evidence that the MD facilitates the routing of OFC inputs to the dlPFC during strategy switches. These findings establish that the MD thalamus serves as a computational hub for regulating frontal cortical interactions in humans. By compressing cortical inputs and providing efficient feedback, the MD enables rapid and flexible gating of dlPFC strategy representations. The study provides direct evidence that the MD integrates inputs from regions like the OFC to dynamically select behavioral strategies, offering a mechanistic explanation for thalamic regulation of cognitive flexibility beyond its classical role as a sensory relay.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-26; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• executive function