No abstract

How does the brain represent and process different types of knowledge? The Dual Hub account postulates that anterior temporal lobes (ATL) support taxonomic relationships based on shared physical features (mole – cat), while temporoparietal regions, including posterior middle temporal gyrus (pMTG), support thematic associations (mole – earth). Conversely, the Controlled Semantic Cognition account proposes that ATL supports both aspects of knowledge, while left pMTG contributes to controlled retrieval. This study used magnetoencephalography to test these contrasting predictions of functional dissociations within the temporal lobe. ATL and pMTG responded more strongly to taxonomic and thematic trials respectively, matched for behavioural performance, in line with predictions of the Dual Hub account. In addition, ATL showed a greater response to strong than weak thematic associations, while pMTG showed the opposite pattern, supporting a key prediction of the Controlled Semantic Cognition account. ATL showed a stronger response for word pairs that were more semantically coherent, either because they shared physical features (in taxonomic trials) or a strong thematic association. These effects largely coincided in time and frequency (although an early oscillatory response in ATL was specific to taxonomic trials). In contrast, pMTG showed non-overlapping effects of semantic control demands and thematic judgements: this site showed a larger oscillatory response to weak associations, when ongoing retrieval needed to be shaped to suit the task demands, and also a larger response to thematic judgements contrasted with taxonomic trials (which was reduced but not eliminated when the thematic trials were easier). Consequently, time-sensitive neuroimaging supports a complex pattern of functional dissociations within the left temporal lobe, which reflects both coherence versus control and distinctive oscillatory responses for taxonomic overlap (in ATL) and thematic relations (in pMTG).

Decomposition of whole-brain functional connectivity patterns reveals a principal gradient that captures the separation of sensorimotor cortex from heteromodal regions in the default mode network (DMN). Functional homotopy is strongest in sensorimotor areas, and weakest in heteromodal cortices, suggesting there may be differences between the left and right hemispheres (LH/RH) in the principal gradient, especially towards its apex. This study characterised hemispheric differences in the position of large-scale cortical networks along the principal gradient, and their functional significance. We collected resting-state fMRI and semantic, working memory and non-verbal reasoning performance in 175 + healthy volunteers. We then extracted the principal gradient of connectivity for each participant, tested which networks showed significant hemispheric differences on the gradient, and regressed participants’ behavioural efficiency in tasks outside the scanner against interhemispheric gradient differences for each network. LH showed a higher overall principal gradient value, consistent with its role in heteromodal semantic cognition. One frontotemporal control subnetwork was linked to individual differences in semantic cognition: when it was nearer heteromodal DMN on the principal gradient in LH, participants showed more efficient semantic retrieval—and this network also showed a strong hemispheric difference in response to semantic demands but not working memory load in a separate study. In contrast, when a dorsal attention subnetwork was closer to the heteromodal end of the principal gradient in RH, participants showed better visual reasoning. Lateralization of function may reflect differences in connectivity between control and heteromodal regions in LH, and attention and visual regions in RH.

We frequently guide our decisions about when and how to act based on the meanings of perceptual inputs: we might avoid treading on a flower, but not on a leaf. However, most research on response inhibition has used simple perceptual stimuli devoid of meaning. In two Go/No-Go experiments, we examined whether the neural mechanisms supporting response inhibition are influenced by the relevance of meaning to the decision, and by presentation modality (whether concepts were presented as words or images). In an on-line fMRI experiment, we found common regions for response inhibition across perceptual and conceptual decisions. These included the bilateral intraparietal sulcus and the right inferior frontal sulcus, whose neural responses have been linked to diverse cognitive demands in previous studies. In addition, we identified a cluster in ventral lateral occipital cortex that was sensitive to the modality of input, with a stronger response to No-Go than Go trials for meaningful images, compared to words with the same semantic content. In a second experiment, using resting-state fMRI, we explored how individual variation in the intrinsic connectivity of these activated regions related to variation in behavioural performance. Participants who showed stronger connectivity between common inhibition regions and limbic areas in medial temporal and subgenual anterior cingulate cortex were better at inhibition when this was driven by the meaning of the items. In addition, regions with a specific role in picture inhibition were more connected to a cluster in the thalamus/caudate for participants who were better at performing the picture task outside of the scanner. Together these studies indicate that the capacity to appropriately withhold action depends on interactions between common control regions, which are important across multiple types of input and decision, and other brain regions linked to specific inputs (i.e., visual features) or representations (e.g., memory).

Semantic control allows us to shape our conceptual retrieval to suit the circumstances in a flexible way. Tasks requiring semantic control activate a large-scale network including left inferior prefrontal gyrus (IFG) and posterior middle temporal gyrus (pMTG) - this network responds when retrieval is focussed on weak as opposed to dominant associations. However, little is known about the biological basis of individual differences in this cognitive capacity: regions that are commonly activated in task-based fMRI may not relate to variation in controlled retrieval. The current study combined analyses of MRI-based cortical thickness with resting-state fMRI connectivity to identify structural markers of individual differences in semantic control. We found that participants who performed relatively well on tests of controlled semantic retrieval showed increased structural covariance between left pMTG and left anterior middle frontal gyrus (aMFG). This pattern of structural covariance was specific to semantic control and did not predict performance when harder non-semantic judgements were contrasted with easier semantic judgements. The intrinsic functional connectivity of these two regions forming a structural covariance network overlapped with previously-described semantic control regions, including bilateral IFG and intraparietal sulcus, and left posterior temporal cortex. These results add to our knowledge of the neural basis of semantic control in three ways: (i) Semantic control performance was predicted by the structural covariance network of left pMTG, a site that is less consistently activated than left IFG across studies. (ii) Our results provide further evidence that semantic control is at least partially separable from domain-general executive control. (iii) More flexible patterns of memory retrieval occurred when pMTG co-varied with distant regions in aMFG, as opposed to nearby visual, temporal or parietal lobe regions, providing further evidence that left prefrontal and posterior temporal areas form a distributed network for semantic control.