Graded modality-specific specialisation in semantics: A computational account of optic aphasia

A long-standing debate regarding the representation of semantic knowledge is whether such knowledge is represented in a single, amodal system or whether it is organised into multiple subsystems based on modality of input or type of information. The current paper presents a distributed connectionist model of semantics that constitutes a middle ground between these unitary- versus multiple-semantics accounts. In the model, semantic representations develop under the pressure of learning to mediate between multiple input and output modalities in performing various tasks. The system has a topographic bias on learning that favours short connections, leading to a graded degree of modality-specific functional specialisation within semantics. The model is applied to the specific empirical phenomena of optic aphasia--a neuropsychological disorder in which patients exhibit a selective deficit in naming visually presented objects that is not attributable to more generalised impairments in object recognition (visual agnosia) or naming (anomia). As a result of the topographic bias in the model, as well as the relative degrees of systematicity among tasks, damage to connections from vision to regions of semantics near phonology impairs visual object naming far more than visual gesturing or tactile naming, as observed in optic aphasia. Moreover, as in optic aphasia, the system is better at generating the name of an action associated with an object than at generating the name of the object itself, because action naming receives interactive support from the activation of action representations. The ability of the model to account for the pattern of performance observed in optic aphasia across the full range of severity of impairment provides support for the claim that semantic representations exhibit graded functional specialisation rather than being entirely amodal or modality-specific.     

Lesion symptom mapping of manipulable object naming in nonfluent aphasia: Can a brain be both embodied and disembodied?

Embodied cognition offers an approach to word meaning firmly grounded in action and perception. A strong prediction of embodied cognition is that sensorimotor simulation is a necessary component of lexical–semantic representation. One semantic distinction where motor imagery is likely to play a key role involves the representation of manufactured artefacts. Many questions remain with respect to the scope of embodied cognition. One dominant unresolved issue is the extent to which motor enactment is necessary for representing and generating words with high motor salience. We investigated lesion correlates of manipulable relative to nonmanipulable name generation (e.g., name a school supply; name a mountain range) in patients with nonfluent aphasia (N?=?14). Lesion volumes within motor (BA4, where BA = Brodmann area) and premotor (BA6) cortices were not predictive of category discrepancies. Lesion symptom mapping linked impairment for manipulable objects to polymodal convergence zones and to projections of the left, primary visual cortex specialized for motion perception (MT/V5+). Lesions to motor and premotor cortex were not predictive of manipulability impairment. This lesion correlation is incompatible with an embodied perspective premised on necessity of motor cortex for the enactment and subsequent production of motor-related words. These findings instead support a graded or “soft” approach to embodied cognition premised on an ancillary role of modality-specific cortical regions in enriching modality-neutral representations. We discuss a dynamic, hybrid approach to the neurobiology of semantic memory integrating both embodied and disembodied components.     

Semantic memory is an amodal,dynamic system: Evidence from the interaction of naming and object use in semantic dementia

Classic neurological accounts and some contemporary theories of semantic memory assume that concepts are acquired through a learning process that draws together information experienced in each of our verbal and nonverbal modalities. These accounts embody three critical assumptions: semantic representations are amodal; the mapping between surface form and meaning varies for different modalities; and the representations are dynamic. The influence of these three factors was revealed in data collected over a 4‐year longitudinal period in two patients with semantic dementia. Semantic assessment revealed a parallel decline in verbal and nonverbal aspects of conceptual knowledge, reflecting a gradual degradation of a single amodal semantic system. As expected, when the patients' semantic impairment was mild, they presented with profound anomia but relatively preserved object use. Over time, performance on all semantic tasks including object use declined. High item‐by‐item consistency across these tasks was observed in all testing sessions. The impact of dynamic semantic representations was revealed by a striking clinical finding. Although unable to name many of the objects in isolation, their performance was significantly facilitated if they were asked to name while they demonstrated the use of each object. These results are discussed in the context of contemporary models of semantic memory.     

Semantic memory is an amodal, dynamic system: Evidence from the interaction of naming and object use in semantic dementia

Classic neurological accounts and some contemporary theories of semantic memory assume that concepts are acquired through a learning process that draws together information experienced in each of our verbal and nonverbal modalities. These accounts embody three critical assumptions: semantic representations are amodal; the mapping between surface form and meaning varies for different modalities; and the representations are dynamic. The influence of these three factors was revealed in data collected over a 4-year longitudinal period in two patients with semantic dementia. Semantic assessment revealed a parallel decline in verbal and nonverbal aspects of conceptual knowledge, reflecting a gradual degradation of a single amodal semantic system. As expected, when the patients' semantic impairment was mild, they presented with profound anomia but relatively preserved object use. Over time, performance on all semantic tasks including object use declined. High item-by-item consistency across these tasks was observed in all testing sessions. The impact of dynamic semantic representations was revealed by a striking clinical finding. Although unable to name many of the objects in isolation, their performance was significantly facilitated if they were asked to name while they demonstrated the use of each object. These results are discussed in the context of contemporary models of semantic memory.     

Word repetition and retrieval practice effects in aphasia: Evidence for use-dependent learning in lexical access

This study tested the hypothesis that a use-dependent learning mechanism operates at each of two stages of lexical access: retrieval of a word from semantics (“Stage 1”), followed by retrieval of the word’s constituent phonemes (“Stage 2”). Two participants with aphasia were selected due to their contrasting types of naming impairment (Stage 1 versus Stage 2 difficulty). For each participant, items were assigned to naming training that involved retrieval practice (retrieval of the name from semantics) or repetition practice (hear the name and orally repeat it). Naming tests were administered one day and one week after training. The results supported the predicted training effects: (a) Because successful naming via retrieval practice requires both Stage 1 and Stage 2, this technique uses and strengthens item-specific connections in both stages. (b) Because word repetition circumvents semantically driven retrieval, this technique primarily uses and strengthens item-specific connections in Stage 2.     

A CASE SERIES ANALYSIS OF “CATEGORY-SPECIFIC” DEFICITS OF LIVING THINGS:THE HIT ACCOUNT

We report a case series analysis of a group of seven patients with apparent “category-specific” disorders affecting living things. On standard diagnostic tests, a range of deficits were apparent, with some cases appearing to have impaired visual access to stored knowledge, some with impaired semantic knowledge (across modalities), and some with an impairment primarily at a name retrieval stage. Patients with a semantic deficit were impaired for both visual and associative/functional knowledge about living things, whilst patients with a modality-specific access deficit showed worse performance when stored visual knowledge was probed. In addition, patients with impaired access to visual knowledge were affected when perceptual input was degraded by masking, and all patients showed an interaction between perceptual similarity and category when matching pictures to names or defining statements. We discuss the results in terms of the Hierarchical Interactive Theory (HIT) of object recognition and naming (Humphreys & Forde, 2001). We also discuss evidence on lesion sites in relation to research from functional brain imaging on category differences in object identification in normal observers.     

A case series analysis of "category-specific" deficits of living things:the hit account

We report a case series analysis of a group of seven patients with apparent "category-specific" disorders affecting living things. On standard diagnostic tests, a range of deficits were apparent, with some cases appearing to have impaired visual access to stored knowledge, some with impaired semantic knowledge (across modalities), and some with an impairment primarily at a name retrieval stage. Patients with a semantic deficit were impaired for both visual and associative/functional knowledge about living things, whilst patients with a modality-specific access deficit showed worse performance when stored visual knowledge was probed. In addition, patients with impaired access to visual knowledge were affected when perceptual input was degraded by masking, and all patients showed an interaction between perceptual similarity and category when matching pictures to names or defining statements. We discuss the results in terms of the Hierarchical Interactive Theory (HIT) of object recognition and naming (Humphreys & Forde, 2001). We also discuss evidence on lesion sites in relation to research from functional brain imaging on category differences in object identification in normal observers.     

Verb production in the nonfluent and semantic variants of primary progressive aphasia: The influence of lexical and semantic factors

Differential patterns of impairment with respect to noun and verb production have been observed in the nonfluent and semantic variants of primary progressive aphasia. However, the factors influencing this discrepancy remain unclear. The present study evaluates verb retrieval in primary progressive aphasia using a naming task and a story completion task. Findings indicate that patients with the semantic variant are influenced by familiarity, frequency, and age of acquisition in both object and action naming, whereas patients with the nonfluent variant are not. Surprisingly, there were no differences in either group between object and action naming, presumably because the lists were well matched on pertinent variables. In the story completion task, greater impairment in semantically heavier than in semantically lighter verbs was observed for the semantic variant, and grammaticality and verb tense agreement was significantly lower in the nonfluent variant. The present findings suggest that lexicosemantic attributes affect verb production in the semantic variant, whereas both lexicosemantic and syntactic attributes affect verb production in the nonfluent variant.     

Action naming with impaired semantics: Neuropsychological evidencecontrasting naming and reading for objects and verbs

We contrast naming from pictures, and reading words, for objects and verbs (actions relating to the objects) in a patient with a large, posterior left-hemisphere lesion. We present evidence for spared picture naming for verbs relative to objects, whilst the opposite pattern of sparing occurred in reading. Objects were also spared relative to verbs in tasks requiring that written words be matched to either pictures or auditory words, in the presence of semantically related or unrelated distractors. We conclude that verb semantics were more impaired than semantic knowledge for objects, and that the better semantic knowledge for object names supported word reading. With pictures, however, action verb retrieval was maintained through a nonsemantic route from vision to action, or though preserved right-hemisphere "action semantics."     

Action naming with impaired semantics: Neuropsychological evidencecontrasting naming and reading for objects and verbs

We contrast naming from pictures, and reading words, for objects and verbs (actions relating to the objects) in a patient with a large, posterior left-hemisphere lesion. We present evidence for spared picture naming for verbs relative to objects, whilst the opposite pattern of sparing occurred in reading. Objects were also spared relative to verbs in tasks requiring that written words be matched to either pictures or auditory words, in the presence of semantically related or unrelated distractors. We conclude that verb semantics were more impaired than semantic knowledge for objects, and that the better semantic knowledge for object names supported word reading. With pictures, however, action verb retrieval was maintained through a nonsemantic route from vision to action, or though preserved right-hemisphere “action semantics.”     

Semantic memory in Alzheimer's disease and the frontotemporal dementias: a longitudinal study of 236 patients

Using semantic dementia (SD) as a reference point, the authors assessed semantic memory in four other neurodegenerative disorders: progressive nonfluent aphasia (PNFA), frontal variant frontotemporal dementia (fvFTD), Alzheimer's disease (AD), and posterior cortical atrophy (PCA). Individuals with SD were more impaired than other groups on semantic measures and showed a characteristic pattern across tasks: category fluency (CF) worse than letter fluency (LF), naming worse than comprehension, and visual and verbal comprehension equally affected, suggesting disruption to an amodal semantic system. Individuals with AD demonstrated a similar pattern to a milder degree. Although PNFA, fvFTD, and PCA groups had abnormal scores (relative to controls) on most semantic measures, their differing patterns across measures indicate that the apparent semantic impairment in these conditions is largely secondary to other factors.     

Neuropsychological evidence for a convergent route model for action

We present a detailed analysis of the relations between acting, naming, and recognising objects from vision, touch, and verbal labels, in a series of three cases. We demonstrate that: (1) there can be better "use" than gesturing of objects even in a patient with impaired semantic information for the objects; (2) this is contingent on direct proprioceptive feedback in action; and (3) there are impaired gestures to "non-action" parts of objects even when the patients can gesture to names and can identify the objects. These results are consistent with a "convergent route" model of action, which holds that direct, modality-specific associations between objects and actions constrain the retrieval of actions from semantic representations.     

Differential effects of advance semantic cues on grasping, naming, and manual estimation

In the three experiments reported here, we systematically investigated when and/or how prior semantic information about a target-object would affect the programming and execution of grasping movements. A name-length association was first created during a practice phase while participants performed one of the three tasks: grasping, naming, or manual estimation of an object’s length upon an auditory presentation of its name. In the subsequent test phase, in the majority of trials, these names correctly indicated the upcoming object for a grasping task or a control naming task (Experiment 1A and 1B), and a grasping task or a manual estimation of object length (Experiment 2). In both experiments, a name was occasionally incongruent or a neutral cue (a burst of white noise) was presented instead. Although in the grasping task, semantic cues (both congruent and incongruent) consistently facilitated movement-onset time (indicating a preparatory-set like effect), the impact of these cues on the formation of grasping aperture was less consistent or absent. The experiments also demonstrated that naming and perceptual estimation of object length were affected differently from grasping by the same semantic cues, with the congruent cues facilitating and the incongruent cues slowing down response onset. These findings reinforce and extend the proposal that the transformations of visual and semantic information for perceptually-driven tasks are quite different from the transformations of the same information for the programming and control of object-directed action.     

ROLE OF MENTAL IMAGERY IN A PROPERTY VERIFICATION TASK: FMRI EVIDENCE FOR PERCEPTUAL REPRESENTATIONS OF CONCEPTUAL KNOWLEDGE

Is our knowledge about the appearance of objects more closely related to verbal thought or to perception? In a behavioural study using a property verification task, Kosslyn (1976) reported that there are both amodal and perceptual representations of concepts, but that amodal representations may be more easily accessed. However, Solomon (1997) argued that due to the nature of Kosslyn's stimuli, subjects may be able to bypass semantics entirely and perform this task using differences in the strength of association between words in true trials (e.g., cat-whiskers) and those in false trials (e.g., mouse-stinger). Solomon found no evidence for amodal representations when the task materials were altered to include associated false trials (e.g., cat-litter), which require semantic processing, as opposed to associative strategies. In the current study, we used fMRI to examine the response of regions of visual association cortex while subjects performed a property verification task with either associated or unassociated false trials. We found reliable activity across subjects within the left fusiform gyrus when true trials were intermixed with associated false trials but not when true trials were intermixed with unassociated false trials. Our data support the idea that conceptual knowledge is organised visually and that it is grounded in the perceptual system.     

Action and object naming in frontotemporal dementia, progressive supranuclear palsy, and corticobasal degeneration

Action naming has been reported to be disproportionately impaired in comparison to object naming in patients with frontotemporal dementia (FTD). This finding has been attributed to the crucial role of frontal cortex in action naming. The investigation of object and action naming in the different subtypes of FTD, as well as in the related conditions of progressive supranuclear palsy (PSP) and corticobasal degeneration (CBD), may thus contribute to the elucidation of the cerebral correlates of the action-object discrepancy as well as provide clues to the underlying cognitive mechanisms. The results indicated that, with the exception of semantic dementia, action naming was more impaired than object naming in all patient groups. The discrepancy was similar in frontal variant of FTD and Alzheimer's disease patients, whereas patients with nonfluent primary progressive aphasia, PSP, and CBD were significantly more impaired in the oral production of actions than of objects. These findings indicate that action naming impairment is not a general feature of FTD, but rather is associated with conditions that affect the frontoparietal-subcortical circuits involved in action knowledge and action representation.     

Role of mental imagery in a property verification task: FMRI evidence for perceptual representations of conceptual knowledge

Is our knowledge about the appearance of objects more closely related to verbal thought or to perception? In a behavioural study using a property verification task, Kosslyn (1976) reported that there are both amodal and perceptual representations of concepts, but that amodal representations may be more easily accessed. However, Solomon (1997) argued that due to the nature of Kosslyn's stimuli, subjects may be able to bypass semantics entirely and perform this task using differences in the strength of association between words in true trials (e.g., cat-whiskers) and those in false trials (e.g., mouse-stinger). Solomon found no evidence for amodal representations when the task materials were altered to include associated false trials (e.g., cat-litter), which require semantic processing, as opposed to associative strategies. In the current study, we used fMRI to examine the response of regions of visual association cortex while subjects performed a property verification task with either associated or unassociated false trials. We found reliable activity across subjects within the left fusiform gyrus when true trials were intermixed with associated false trials but not when true trials were intermixed with unassociated false trials. Our data support the idea that conceptual knowledge is organised visually and that it is grounded in the perceptual system.     

The role of action representations in visual object recognition

It is typically assumed that perception for action and object recognition are subserved by functionally and neuroanatomically distinct processing streams in the brain. However, recent evidence challenges this classical view and suggests an interaction between both visual processing streams. While previous studies showed an influence of object perception on action-related tasks, we investigated whether action representations facilitate visual object recognition. In order to address this question, two briefly displayed masked objects were sequentially presented, either affording congruent or incongruent motor interactions. We found superior naming accuracy for object pairs with congruent as compared to incongruent motor interactions (Experiment 1). This action priming effect indicates that action representations can facilitate object recognition. We further investigated the nature of the representations underlying this action priming effect. The effect was absent when the prime stimulus was presented as a word (Experiment 2). Thus, the action priming effect seems to rely on action representations specified by visual object information. Our findings suggest that processes of object-directed action influence object recognition.     

Agnosia for global patterns: When the cross-talk between grouping and visual selective attention failS

In this paper we consider whether our stored conceptual knowledge about stimuli is represented within a single semantic system which is indifferent to the modality of stimulus presentation, or whether conceptual knowledge is represented in different semantic systems according to either the modality of stimulus presentation or the nature of the concept (e.g. whether the concept specifies visual or verbal knowledge about an object). Previous work suggests three areas of neuropsychological research which are relevant to this issue: modality-specific aphasias, modality-specific priming on semantic access disorders, and modality-specific aspects of semantic memory disorders. Evidence from each of these areas is reviewed and we argue that there is only equivocal support for the multiple semantic systems position. We outline an alternative account which distinguishes between a single a modal semantic system and modality-specific perceptual recognition systems, and we discuss the evidence in light of this single semantic system account.     

Visual object processing in optic aphasia: A case of semantic access agnosia

A single case study is reported of a patient with a naming disorder specific to visually presented stimuli. The patient was often able to gesture correctly to objects he could not name, and he showed intact access to structural knowledge of objects.

Further examination revealed an impairment in accessing semantic knowledge about objects, which was most marked when the patient had to discriminate between objects which were visually as well as semantically similar. It is suggested that the patient's naming deficit is due to an impairment in accessing semantic information from vision, following intact access to stored structural knowledge. Correct gestures may be contingent on access to the system specifying structural knowledge. The data are interpreted in terms of a model of visual object identification in which access to semantic information, from the system specifying structural knowledge, is held to operate in cascade.