Research ReportIntrinsic connectivity of anterior temporal lobe relates to individual differences in semantic retrieval for landmarks
Introduction
Semantic cognition allows us to understand the world around us–including the meaning of words, objects, locations and people (Lambon Ralph et al., 2017; Patterson et al., 2007). Conceptual representations that underpin semantic performance across input modalities (e.g., words and pictures) and across different tasks are thought to be supported by the bilateral ventral anterior temporal lobes (ATL; Binney et al., 2010; Rice, Hoffman, & Lambon Ralph, 2015, 2015b). Patients with semantic dementia have degraded conceptual knowledge that is associated with bilateral ATL atrophy, while other aspects of cognition remain largely intact (Lambon Ralph et al., 2017; Patterson et al., 2007). Semantic deficits following ATL damage are most pronounced for semantic tasks that probe specific-level knowledge–including knowledge of unique entities including people, such as Barack Obama, and landmarks, such as the Eiffel Tower (Rogers et al., 2015; Tranel, 2006; Tranel et al., 1997). An influential account of ATL function suggests that this region forms a semantic “hub” drawing together different features represented within ‘spokes’ (capturing visual, valence, language and auditory inputs) to form heteromodal concepts (Patterson et al., 2007). More detailed conceptual retrieval may be needed to identify unique entities and other specific concepts within this hub (Rogers et al., 2006); in line with this, unique entities activate ATL more strongly than more general-level concepts in neuroimaging studies of healthy participants (Gorno-Tempini & Price, 2001; Grabowski et al., 2001; Ross & Olson, 2012).
Neuropsychological and neuroimaging studies suggest that semantic representations draw on a bilateral hub, implemented across both left and right ATL (Ding et al., 2020; Lambon Ralph et al., 2017; Patterson et al., 2007). Severe semantic degradation follows bilateral ATL atrophy in semantic dementia; in contrast, patients with unilateral lesions following resection for temporal lobe epilepsy have measurable yet much milder semantic deficits (Rice, Caswell, et al., 2018). This might reflect functional compensation by the intact ATL (Jung & Lambon Ralph, 2016). Neuroimaging studies with healthy participants have observed bilateral responses to semantic tasks in ATL (Visser et al., 2009), irrespective of whether words or pictures are presented (Bright et al., 2004; Tranel et al., 2005; Vandenberghe et al., 1996; Visser et al., 2012; Visser & Lambon Ralph, 2011). Inhibitory transcranial magnetic stimulation (TMS) delivered to either left or right ATL disrupts both picture and word-based semantic tasks, mimicking the pattern in semantic dementia (Pobric et al, 2007, 2010). Moreover, in line with expectations for a single semantic hub distributed across two hemispheres (cf. Schapiro et al., 2013), inhibitory TMS to left ATL leads to an increase in the response within right ATL, suggesting the non-stimulated hemisphere may compensate for functional disruption within the stimulated hemisphere (Binney & Lambon Ralph, 2015; Jung & Lambon Ralph, 2016).
There is also evidence that left and right ATL are not functionally identical. The integration of different aspects of knowledge in ATL is thought to occur in a graded connectivity-dependent fashion (Bajada et al., 2019; Binney et al., 2012), with the most heteromodal responses in ventrolateral ATL (Lambon Ralph et al., 2017; Murphy et al., 2017; Visser et al., 2012). A degree of functional specialisation across the hemispheres might reflect differential white matter connections from visual, auditory-motor, social and emotional networks to left and right ATL (Papinutto et al., 2016; Rice, Hoffman, & Lambon Ralph, 2015). Some studies have suggested a modality difference across the hemispheres–with left ATL showing stronger engagement for verbal tasks, and right ATL showing a preference for non-verbal tasks. Patients with semantic dementia who have relatively more left-sided or right-sided atrophy provide support for this suggestion (Gainotti, 2012): atrophy in right ATL correlates with difficulties on picture semantic tasks, while damage to left ATL is more strongly correlated with verbal semantic task performance (Butler et al., 2009; Mion et al., 2010). A variant of this modality view suggests that output modality is also important–damage to left ATL is particularly associated with problems in naming concepts (Lambon Ralph et al., 2001), and therefore with deficient lexical access from semantic knowledge, while right ATL is linked to poor object recognition (Damasio et al., 2004). These effects of modality across left and right ATL are most commonly reported in studies comparing semantic retrieval to people's names and faces (Gainotti, 2007, 2013; Gainotti & Marra, 2011; Luzzi et al., 2017; Snowden et al, 2012, 2017): for example, Snowden et al. (2004) found that patients with more left-lateralised atrophy had greater impairment for people's names, while patients with more right-sided atrophy had greater difficulty on semantic tasks employing faces. Given that multiple neuropsychological studies have found hemispheric differences in ATL when contrasting famous people's names and faces, individual differences in connectivity from left and right ATL in healthy participants might also relate to these aspects of semantic cognition (see below; less is currently known about modality effects for the landmark category).
An alternative account suggests that hemispheric specialisation in left and right ATL reflects semantic category, not (only) input modality–by this view, right ATL has been argued to play a larger role than left ATL in understanding social concepts and retrieving conceptual information about specific people (Zahn et al., 2007). Patients with damage to right ATL often have particular difficulties recognising faces, but there is an ongoing debate about whether these difficulties reflect impairment for faces per se (i.e., difficulty when the task involves both social stimuli and picture inputs) or a wider problem with social concepts (Gainotti, 2013; Gorno-Tempini et al., 1998; Olson et al., 2007, 2013; Ross & Olson, 2010). In addition, a few neuroimaging studies have contrasted knowledge of famous faces and names with knowledge of landmarks (since both of these categories comprise unique entities): this contrast reveals common activation for both faces and landmarks in left ATL (Gorno-Tempini & Price, 2001; Grabowski et al., 2001; Wang et al., 2016), and weaker engagement of right ATL across both categories. A recent fMRI study (Rice, Hoffman, et al., 2018) directly compared the neural response in ATL during semantic decisions about specific entities that were social (people) and non-social (landmarks). The social and non-social stimuli were presented as both words (i.e., people's names) and as pictures (i.e., faces). Both the left and right ventral ATL responded regardless of the modality or category of semantic information, although an additional bilateral region in the ATL, extending towards the temporal pole, showed stronger activation to people versus landmarks. Similarly, a meta-analysis of 97 functional neuroimaging studies (Rice, Lambon Ralph, & Hoffman, 2015) confirmed that both left and right ATL were activated across verbal and non-verbal stimuli, and social and non-social tasks. Studies involving word retrieval were more likely to report unilateral left ATL activation, while social semantic studies were more likely to observe bilateral ATL activation.
Rice and colleagues (2015b) suggested that subtle functional differences between left and right ATL are likely to reflect differential connectivity to ‘spoke’ systems–for example, stronger connectivity between right ATL and visual regions, or between left ATL and language regions. In this extension of the “graded hub account”, the bilateral semantic store shows some degree of differentiation between left and right ATL, such that distinct patterns of connectivity can explain modality effects (for example, differences between faces and names). Semantic retrieval for different categories is also thought to draw disproportionately on particular feature types–for example, visual textures and movement are critical to the animal category, while action features and hand shape are important for tools (Fernandino et al., 2016; Ishibashi et al., 2016; Liljeström et al., 2008; Mollo et al., 2018; Moss et al., 1998, 2005; Rogers et al., 2005). Consequently, this type of account is potentially able to accommodate differences in the contributions of left and right ATL to distinct semantic categories, such as people versus landmarks, as well as effects of modality. In a recent study (Gonzalez Alam et al., 2019), left ATL was more connected with other sites implicated in semantic cognition (e.g., left inferior frontal gyrus, posterior middle and inferior temporal cortex), while right ATL was more connected to visual cortex and default mode network regions, including angular gyrus and dorsomedial prefrontal cortex. Despite these differences, ATL had the most symmetrical connections within the semantic network (i.e., the highest correlations between connectivity patterns generated from left and right-hemisphere seeds), consistent with the bilateral hub theory.
Individual differences in patterns of connectivity from left and right ATL might differentially relate to the efficiency of semantic decisions about people and landmarks, presented as words and pictures. Gonzalez Alam et al. (2019) failed to observe any behavioural correlates of hemispheric differences in ATL connectivity; however, this previous study did not use tasks designed to maximise the involvement of this brain area (i.e., specific-level concepts). Studies show that individual differences between participants in the strength of intrinsic connectivity, defined according to time–series correlations in resting-state fMRI, are associated with variation in cognitive performance (Cole et al., 2014; Smith et al., 2009; Van Dijk et al., 2010), including semantic cognition (Evans et al., 2020; Gonzalez Alam et al., 2018; Mollo et al., 2016; Poerio et al., 2017; Vatansever et al., 2017; Zhang et al., 2019). Hemispheric differences in intrinsic connectivity across individuals have already been linked to semantic and spatial tasks (Gonzalez Alam et al., 2019; Sormaz et al., 2017; although these prior studies did not find associations with ATL connectivity). Individual differences in the connectivity of ATL have also been associated with semantic processing, although these studies did not examine hemispheric differences (Mollo et al., 2016; Vatansever et al., 2017). To address this gap in the literature, the current study acquired resting-state fMRI from 74 participants, who completed semantic decisions about specific people and landmarks, presented as written words and pictures (using the stimuli from Rice, Hoffman, et al., 2018) in a separate testing session following the scan. We then assessed relationships between connectivity and behavioural performance–with particular focus on whether right versus left ATL connectivity would predict performance on these different categories presented as pictures versus words.
Section snippets
Methods
This study includes analysis of intrinsic connectivity in a sample of participants who performed a resting-state scan and behavioural testing in two separate sessions. We determined our sample size based on participant availability (i.e., tested as many of the participants with resting-state data as were willing), and report all data exclusions, all inclusion/exclusion criteria, whether inclusion/exclusion criteria were established prior to data analysis, all manipulations, and all measures in
Behavioural results
Fig. 2 shows the average median reaction time, accuracy and efficiency scores in each condition of the semantic task adapted from Rice, Hoffman, et al. (2018). A two-way repeated measures ANOVA on reaction time data with category and modality as factors revealed no main effect of modality, a significant main effect of category and a category by modality interaction [Category: F (2,144) = 56.07, p < .001; Interaction: F (2,144) = 69.1, p < .001]. Post-hoc tests with Bonferroni correction showed
Discussion
This study examined the relationship between individual differences in the intrinsic connectivity of left and right ATL and performance on semantic tasks that involved (i) different modalities of presentation (pictures vs words) and (ii) knowledge of specific people versus landmarks. Previous work has suggested that while the functions of left and right ATL are more similar than they are different, there is some subtle hemispheric specialisation according to modality–in particular, patients
Author contributions
Tirso RJ Gonzalez Alam: Conceptualization, Methodology, Software, Formal analysis, Investigation, Data curation, Visualization, Writing- Original draft preparation, Writing- Reviewing and Editing.
Katya Krieger-Redwood: Methodology, Formal analysis, Investigation.
Megan Evans: Formal analysis, Investigation.
Grace E Rice: Software, Writing- Reviewing and Editing.
Jonathan Smallwood: Conceptualization, Methodology, Formal analysis, Resources, Visualization, Writing- Original draft preparation,
Funding
EJ was supported by European Research Council [FLEXSEM-771863], JS was supported by European Research Council [WANDERINGMINDS-646927], TGA was supported by the National Council of Science and Technology, Mexico [Scholarship 411361].
Open Practices
The study in this article earned an Open Materials badge for transparent practices. Materials and data for the study are available at https://neurovault.org/collections/5687/.
Declaration of competing interest
None.
Acknowledgements
We would like to thank the students that helped with data collection: Melinda Bjolseth, Sophie Colgan, Lucy Milne, Clara Scatola, Nikita Schaap, George Shone, Molly Sibson-Flood, Dragos Teodorov and Molly Wilson.
References (97)
- et al.
A structural connectivity convergence zone in the ventral and anterior temporal lobes: Data-driven evidence from structural imaging
Cortex; a Journal Devoted To the Study of the Nervous System and Behavior
(2019) - et al.
Grounding conceptual knowledge in modality-specific systems
Trends Cognate Science
(2003) - et al.
A component based noise correction method (CompCor) for BOLD and perfusion based fMRI
Neuroimage
(2007) - et al.
Using a combination of fMRI and anterior temporal lobe rTMS to measure intrinsic and induced activation changes across the semantic cognition network
Neuropsychologia
(2015) - et al.
Unitary versus multiple semantics: PET studies of word and picture processing
Brain and Language
(2004) - et al.
Anticorrelations in resting state networks without global signal regression
Neuroimage
(2012) - et al.
Controlled semantic cognition relies upon dynamic and flexible interactions between the executive ‘semantic control’ and hub-and-spoke ‘semantic representation’ systems
Cortex; a Journal Devoted To the Study of the Nervous System and Behavior
(2018) - et al.
Intrinsic and task-evoked network architectures of the human brain
Neuron
(2014) - et al.
Neural systems behind word and concept retrieval
Cognition
(2004) - et al.
Controlled semantic summation correlates with intrinsic connectivity between default mode and control networks
Cortex; a Journal Devoted To the Study of the Nervous System and Behavior
(2020)
Category-selective neural substrates for person- and place-related concepts
Cortex; a Journal Devoted To the Study of the Nervous System and Behavior
Different patterns of famous people recognition disorders in patients with right and left anterior temporal lesions: A systematic review
Neuropsychologia
The format of conceptual representations disrupted in semantic dementia: A position paper
Cortex; a Journal Devoted To the Study of the Nervous System and Behavior
Degrees of lateralisation in semantic cognition: Evidence from intrinsic connectivity
Neuroimage
Meaningful inhibition: Exploring the role of meaning and modality in response inhibition
Neuroimage
The neural basis of object perception
Current Opinion in Neurobiology
Perceiving and naming actions and objects
Neuroimage
Famous faces and voices: Differential profiles in early right and left semantic dementia and in Alzheimer's disease
Neuropsychologia
Coming of age: A review of embodiment and the neuroscience of semantics
Cortex; a Journal Devoted To the Study of the Nervous System and Behavior
Context-dependent lexical ambiguity resolution: MEG evidence for the time-course of activity in left inferior frontal gyrus and posterior middle temporal gyrus
Brain and Language
An individual differences analysis of the neurocognitive architecture of the semantic system at rest
Brain and Cognition
The impact of global signal regression on resting state correlations: Are anti-correlated networks introduced?
Neuroimage
Fractionating the anterior temporal lobe: MVPA reveals differential responses to input and conceptual modality
Neuroimage
Amodal semantic representations depend on both anterior temporal lobes: Evidence from repetitive transcranial magnetic stimulation
Neuropsychologia
Disorders of representation and control in semantic cognition: Effects of familiarity, typicality, and specificity
Neuropsychologia
Social cognition and the anterior temporal lobes
Neuroimage
Knowing what from where: Hippocampal connectivity with temporoparietal cortex at rest is linked to individual differences in semantic and topographic memory
Neuroimage
The neuroscience of remote spatial memory: A tale of two cities
Neuroscience
A neural basis for the retrieval of conceptual knowledge
Neuropsychologia
Varieties of semantic cognition revealed through simultaneous decomposition of intrinsic brain connectivity and behaviour
Neuroimage
The role of vision in the neural representation of unique entities
Neuropsychologia
Origins of landmark encoding in the brain
Trends in Neurosciences
Topographical disorientation: A synthesis and taxonomy
Brain: a Journal of Neurology
Perceptual symbol systems
The Behavioral and Brain Sciences
The ventral and inferolateral aspects of the anterior temporal lobe are crucial in semantic memory: Evidence from a novel direct comparison of distortion-corrected fMRI, rTMS, and semantic dementia
Cerebral Cortex
Convergent connectivity and graded specialization in the rostral human temporal lobe as revealed by diffusion-weighted imaging probabilistic tractography
Journal of Cognitive Neuroscience
The neural correlates of verbal and nonverbal semantic processing deficits in neurodegenerative disease
Cognitive and Behavioral Neurology: Official Journal of the Society for Behavioral and Cognitive Neurology
Unveiling the dynamic interplay between the hub- and spoke-components of the brain's semantic system and its impact on human behaviour
From perception to conception: How meaningful objects are processed over time
Cerebral Cortex
The evolution of meaning: Spatio-temporal dynamics of visual object recognition
Journal of Cognitive Neuroscience
A unified neurocognitive model of semantics language social behaviour and face recognition in semantic dementia
Nature Communications
Combining reaction time and accuracy: The relationship between working memory capacity and task switching as a case example
Perspectives on Psychological Science
Cluster failure: Why fMRI inferences for spatial extent have inflated false-positive rates
Perspectives on Psychological Science
Person- and place-selective neural substrates for entity-specific semantic access
Cerebral Cortex
Concept representation reflects multimodal abstraction: A framework for embodied semantics
Cerebral Cortex
Is the right anterior temporal variant of prosopagnosia a form of “associative prosopagnosia” or a form of “multimodal person recognition disorder”?
Neuropsychology Review
Differential contribution of right and left temporo-occipital and anterior temporal lesions to face recognition disorders
Frontiers in Human Neuroinformatics
The fusiform “face area” is part of a network that processes faces at the individual level
Journal of Cognitive Neuroscience
Cited by (7)
Individual differences in gradients of intrinsic connectivity within the semantic network relate to distinct aspects of semantic cognition
2022, CortexCitation Excerpt :In order to further investigate this possibility, we extracted intrinsic connectivity between (i) key nodes within the semantic control network (i.e., left IFG and pMTG) and (ii) semantic regions not associated with control processes (left ATL and AG). The coordinates for the seeds were used in a recent study of intrinsic connectivity within the semantic network (Gonzalez Alam et al., 2021a) and were derived from meta-analyses of (i) semantic control, for IFG (−48 22 20) and pMTG (−54 −42 4; Jackson, 2020), (ii) semantic processing within ATL (−41 −15 −31; averaged from peak coordinates of semantic > non-semantic contrasts listed in Table 1 of Rice, Hoffman, Binney, & Lambon Ralph, 2018) and (iii) AG (−48 −64 34; Humphreys & Lambon Ralph, 2015). Although left AG is functionally heterogeneous and its contribution to semantic cognition remains controversial, recent studies have suggested left mid-AG is associated with more automatic aspects of semantic retrieval, as well as episodic memory (e.g., Humphreys & Lambon Ralph, 2015; Humphreys, Lambon Ralph, & Simons, 2021).
Task modulation of spatiotemporal dynamics in semantic brain networks: An EEG/MEG study
2022, NeuroImageCitation Excerpt :Few studies have investigated the interaction between semantic control and representation regions, and the connectivity and temporal dynamics of the corresponding brain regions are still not well understood (Jefferies, 2013; Lambon Ralph et al., 2016). Most previous studies of the semantic network and its connectivity have employed fMRI (Alam et al., 2021; Chiou et al., 2018; Chiou and Lambon Ralph, 2019; Humphreys et al., 2015; Jackson et al., 2016; Kuhnke et al., 2020) which despite its excellent spatial resolution, is limited in tracking any neural response faster than one second. In the present study, we provide novel evidence for task effects on the spatiotemporal dynamics in the semantic network and, in particular, functional connectivity amongst these regions.
A semantic retrieval model of social media data based on statistical theory
2024, International Journal of Web Based CommunitiesA tale of two gradients: differences between the left and right hemispheres predict semantic cognition
2022, Brain Structure and Function