Week 2 RF-Semantic Cognition Flashcards
What are the 2 key components in semantic cognition?
- Representation:
-Anterior temporal lobe reflects our database (i.e., storage & representations) of semantic knowledge
-If this is lesioned, we do have a multimodal central impairment in semantic processing. SD (semantic dementia) patients are not able to tell the meaning of a picture of a piano or the word itself, and cannot tell where the sound comes from (i.e., no semantic judgement on stimuli presented) - Control:
-Inferior Frontal Gyrus,
-Inferior Parietal lobe / Intra-parietal sulcus
-Posterior Middle Temporal Gyrus reflects processes for flexible use of semantic knowledge
-This network is important to select the semantic information that is relevant for the task at hand. So, for instance, if we want to name a picture of a piano, we will need to select the relevant representations to perform that task, that is lexical and phonological representations.
-Patients who have lesions in this network have issues of access – that is the representation is still there but there is an impairment of the control system that allows for retrieving the correct representations. So in naming for instance these patients can do the naming task if there are facilitated with phonological cues.
What is the Controlled Semantic Cognition (CSC) framework? (Lambon Ralph et al., 2017)
-This framework relates to how concepts are activated and represented in the brain.
-So the key idea is that concept needs requires activation across all these information specific areas combined with the ATL which acts as a sort of orchestra conductor who brings together all the relevant and required bits of information for any specific concepts.
-Primary modal and association modal regions = spokes
-The spokes are joined by the ATL. This works with the distributed information-specific regions to form a coherent concept by combining together all the information and re-representing it in an abstract, a-modal form.
-Hetero-modal association region = a single hub
-Heteromodal (polymodal) association cortex: This area integrates information from multiple sensory modalities. It plays a critical role in higher cognitive functions, including attention, memory, language, and conscious thought.
-Unimodal association cortex: This area receives input from a single sensory modality. The output from the unimodal cortex is largely to unimodal areas. For example, the visual association cortex in the occipital lobe processes visual information, and the auditory association cortex in the temporal lobe processes auditory information.
-This idea of the spokes and hub model has been also implemented formally in a computational model.
What was Binder et al’s (2009) study?
-Meta-analysis of neuroimaging studies
-PET or fMRI study using words
-Task contrast targeting concept retrieval processes
-Controls for orthographic and phonological demands
-Controls for executive and cognitive control demands
-Important in these 120 studies the authors were not looking at things like sound or colour they were just looking at meaningful versus not meaningful stimuli.
What was the findings involved in Binder et al’s (2009) study?
-1145 activation peaks from 120 studies, projected to a brain template surface (i.e., across brain regions)
-As you can see besides the ATL we have also other brain regions and it seems that a lot of these activation peaks fall into the IPL – which in the previous model was not mentioned as being one important hub for semantics.
What is the Multi-hobs model? (Binder et al., 2009; Binder & Desai, 2011)
-High-level, highly convergent and hetero-modal hubs
-Extra areas of processing not accounted for in the prior model e.g., the AG or the posterior cingulate cortex
Multi-hub models: What do the Primary modal and association modal regions do? (Binder et al., 2009; Binder & Desai, 2011)
They process information that is modality-specific (sensory, action, and emotion systems).
They also provide experiential input to high-level
Multi-hub models: What do the Hetero-modal convergence zones do? (Binder et al., 2009; Binder & Desai, 2011)
They combine information across different modalities and store increasingly abstract representations of entity and event knowledge.
They also store increasingly abstract representations of entity and event knowledge.
What evidence is there supporting the role of the Angular gyrus (AG) in semantic cognition?
-fMRI studies reveal consistent AG engagement for general semantic contrasts (e.g., words vs. pseudowords; Binder et al. 2009; Wirth et al., 2011; Jackson 2021).
-TMS studies indicate a causal role of left AG in general semantic processing (Sliwinska 2015; Davey et al., 2015).
-Many studies have suggested that processing meaningful vs non-meaningful items results in the activation of the AG.
-But we also have TMS results that suggest that AG has a key role for semantic processing.
-TMS is important because it demonstrates a causal role for a specific brain region.
What are Geschwind’s language models? (Geschwind 1968, 1972)
-Relatively limited considerations of language comprehension (the model contains only the dorsal pathway)
-Considered the angular gyrus to be a key region
-AG is multimodally connected and thus an obvious region for merging information into concepts
-‘When lesioned the AG generates selective semantic impairments – i.e., without affecting other aspects of language (Geschwind et al., 1968)’
-This view has been highly influential and still is a dominant axiom in contemporary research
-Geschwind reports this case of a young woman who following an anoxic episode was found to have severe semantic deficits whilst retaining other key language functions including repetition, nonpropositional speech and the ability to learn new songs (transcortical sensory aphasia: Geschwind et al. 1968). The case subsequently came to autopsy and thus Geschwind et al. were able to infer functional localization (this was, of course, long before brain scanning was possible). As noted by Geschwind, the patient’s watershed lesion had damaged the AG.
-The AG has long been strongly associated with semantic memory, stemming back at least as far as Geschwind’s seminal studies (Geschwind 1972).
-However, with the benefit of contemporary findings from neuropsychology and functional neuroimaging, one of the most striking aspects of revisiting the pathological report is the finding of severe damage in all other parts of the semantic network including bilateral ventral ATL, MTG, and prefrontal cortex amongst other regions (Geschwind et al., 1968).
What is one of the controversies? (Humphreys et al., 2015)
-Response polarity – AG was similarly deactivated in both semantic and non-semantic tasks
-According to this “semantic hypothesis” (Geschwind 1972; Binder et al. 2009), the AG is a hub of multimodal conceptual knowledge or a hub for certain types of multimodal semantic dimensions which preferentially responds to semantic information (Not seen in results above)
What is one of the controversies? (Humphreys et al., 2017)
-AG’s responses and task difficulty – instead of responding to semantic manipulations (meaningful vs non-meaningful), AG responds to task-difficulty manipulations
What is one of the controversies? (Jefferies & Lambon Ralph, 2006)
-Posterior semantic aphasic patients (whose lesions overlap with AG) do not seem to have degraded semantic representations, unlike those with anterior temporal damage, such as semantic dementia
What are some interesting facts relating to the AG?
fMRI studies show that AG is activated by:
-Semantic manipulations, but only when the task requires multi-item processing (Branzi et al., 2020; Humphreys et al., 2020)
-Language production and language comprehension (Silbert et al., 2014)
-Non-semantic tasks (Humphreys et al., 2015; 2020) – e.g., number sequence processing, episodic memory, music processing…
THEREFORE does AG support a buffering system?
What was the reading task completed on online TMS? (Branzi et al., 2020, 2021)
Context:
We learnt many things that day. We learnt about elephants and the natural disaster in Tanzania. Indeed, two major climatic changes had drastically affected the number of elephants.
Target:
We also learnt that trunks can be used by elephants to rub an itchy eye. Moreover, these animals use their trunk to threaten and to throw objects, and as snorkels when swimming in water.
-Asked to read either narrative and understand the meaning to then be asked some questions
-TMS protocol: Train of 5 pulses (10hz, 500ms) [e.g., Sliwinska et al., 2015] delivered online at 120% of the RMT between prime and target trials.
-Stimulation sites: left AG (group coordinate determined from the fMRI study) and Vertex (control site).
What was the memory task with no TMS? (Branzi et al., 2020, 2021)
Prediction: TMS on AG affects memory retrieval of contextual information
-The memory task requires having formed a ‘context and target integrated representation’ during the reading task
-By using TMS, there should be a disruption seen in performance
