Where is the semantic network?. What is the semantic network? Knowledge of objects, people, concepts and word meanings Spreading-activation theory.

Where is the semantic network?

What is the semantic network?

Knowledge of objects, people, concepts and word meanings

Spreading-activation theory of semantic processing (Quillian, 1962; Collins & Loftus, 1975)

Other models also theorise how the knowledge is represented within the mental lexicon; along with phonological and syntactic information (Roelofs, 1992; Bock & Levelt, 1994)

What is the semantic network?

Knowledge of objects, people, concepts and word meanings

Spreading-activation theory of semantic processing (Quillian, 1962; Collins & Loftus, 1975)

Other models also theorise how the knowledge is represented within the mental lexicon; along with phonological and syntactic information (Roelofs, 1992; Bock & Levelt, 1994)

Where is the semantic network?

Impairments of semantic processing (frontotemporal dementia, aphasia, Alzheimer's disease, autism and schizophrenia)

Broadly distributed neural representation, with marked reliance on inferotemporal as well as posterior inferior parietal regions (e.g., Damasio et al., 2004)

All these brain regions play a role in high-level interactive processes; they receive extensively processed, multimodal and supramodal input

Neuroimaging findings

120 fMRI studies (Binder et al., 2009) Left-lateralised seven-region network:

Posterior inferior parietal lobe (AG, SMG) Lateral temporal cortex (MTG, ITG) Ventral temporal cortex (PH, FFG) Ventromedial prefrontal cortex Dorsomedial prefrontal cortex Posterior cingulate gyrus Inferior frontal gyrus

PET and fMRI studies (Price, 2012)

Research area

Within-system specialisation: distinct tasks, contrasts, processes; some subsystems may specialise in specific object categories, attributes, or type of knowledge

Patients with profound object recognition disorders have intact word comprehension; no/minimal overlap between the systems underlying word and object recognition (Warrington, 1985; Farah, 1990)

fMRI studies support the view that comprehension of a word does not activate a perceptual representation of the object to which it refers (e.g., Moore & Price, 1999)

Neuropsychological evidence

Semantic dementia Symptomatology:

Progressive loss of conceptual knowledge that leads to anomia, impaired comprehension and semantically invalid speech (Davies et al., 2005)

Neuropathology: Typically left-sided atrophy in the temporal

lobe with a marked anterior gradient Marked loci are: temporal pole, middle and

inferior temporal gyri, and fusiform gyrus (Chan et al., 2000; Mummery et al. 2000)

Neuropsychological evidence

Semantic aphasia Symptomatology:

Impaired semantic retrieval and control but intact knowledge of meanings when there are no executive control demands (e.g., Robinson et al., 2005; Novick et al., 2009)

Neuropathology: Post-stroke lesions to the frontal (IFG)

and/or temporoparietal cortices (e.g., Wagner et al., 2001; Badre, 2008)

Double dissociation

Understanding word meanings relies on two processes:

Activation of word meanings Retrieval and manipulation of the information

in a given situation/task

These two processes of semantic cognition seem to be computed by distinct brain areas:

Representation – aITG Control/retrieval – pMTG, IFG

Semantic representation versus control

Whitney, Jefferies, & Kircher (2011)

Semantic relatedness judgement task

Semantic representation versus control

Whitney, Jefferies, & Kircher (2011)

Semantic relatedness judgement task

LION

Semantic representation versus control

Whitney, Jefferies, & Kircher (2011)

Semantic relatedness judgement task

LION STRIPE

Semantic representation versus control

Whitney, Jefferies, & Kircher (2011)

Semantic relatedness judgement task

LION STRIPE TIGER

Semantic representation versus control

Whitney, Jefferies, & Kircher (2011)

Semantic relatedness judgement task

LION STRIPE TIGER

Is this word related to any of the last two

words?YES NO

Semantic representation versus control

Prime-target relationships:

1. Ambiguous double-related (game-dance-ball)

2. Unambiguous double-related (lion-stripe-tiger)

3. Single-related superordinate (game-pillow-ball)

4. Single-related subordinate (dance-clock-ball)

Ambiguous > unambiguous trials

Ambiguous > unambiguous trials

High > low semantic control demands

Proposed research

To identify the brain region (within the temporal lobe) that stores semantic representations

Three-phase research project:

1. Acquisition of novel word meanings2. Mapping semantic representations (fMRI)3. Validation (TMS)

Acquisition of novel words

40 novel words (e.g., freckton) Word forms and meanings learnt from context

within 100-word paragraphs Word production task 5 online worksheets

Lexical decision task: Have the new word forms been integrated?

Novel words ~ low-frequency long words Novel words > pseudo-words

Acquisition of novel words

Masked semantic priming task: Have the new meanings been fully

consolidated? Presence of the priming effect

#######

Acquisition of novel words

Masked semantic priming task: Have the new meanings been fully

consolidated? Presence of the priming effect

FRECKTON#######

Acquisition of novel words

Masked semantic priming task: Have the new meanings been fully

consolidated? Presence of the priming effect

FRECKTON####### MOUSE

Acquisition of novel words

Masked semantic priming task: Have the new meanings been fully

consolidated? Presence of the priming effect

FRECKTON####### MOUSE

DOG

WHISKERS

Mapping semantic representations (fMRI) Following successful consolidation of both word

forms and meanings (indexed by ad hoc latency data cut-off)

A number of different tasks that do not necessitate semantic control demands or executive processes (e.g., silent word reading paradigm)

Subjects’ attention maintained with inter-trial reward

Multiple target word presentation

Mapping semantic representations (fMRI) Mixed/blocked design

Comparison of the activation loci pre and post the acquisition of novel words

ROI analysis (temporal lobe)

Validation of the findings (TMS)

Stimulation over the region of interest (aITG)

Behavioural data: Performance on semantic access tasks with

no/minimal semantic control demands e.g., Word naming task Attribute recall task Meaning retrieval task