Structural similarity and category-specificity: a refined account

Neuropsychological evidence for the temporal dynamics of category-specific naming

Multiple accounts have been proposed to explain category-specific recognition impairments. Some suggest that category-specific deficits may be caused by a deficit in recurrent processing between the levels of a hierarchically organized visual object recognition system. Here, we tested predictions of interactive processing theories on the emergence of category-selective naming deficits in neurologically intact observers and in patient GA, a single case showing a category-specific impairment for natural objects after a herpes simplex encephalitis infection. Fragmented object outlines were repeatedly presented until correct naming occurred (maximum 10 times), and the fragments increased in length with every repetition. We studied how shape complexity, object category, and fragment curvature influence the timing of correct object identification. The results of a survival analysis are consistent with the idea that deficits in recurrent processing between low- and high-level visual object representations can cause category-selective impairments.

A unified model of human semantic knowledge and its disorders

How is knowledge about the meanings of words and objects represented in the human brain? Current theories embrace two radically different proposals: either distinct cortical systems have evolved to represent different kinds of things, or knowledge for all kinds is encoded within a single domain-general network. Neither view explains the full scope of relevant evidence from neuroimaging and neuropsychology. Here we propose that graded category-specificity emerges in some components of the semantic network through joint effects of learning and network connectivity. We test the proposal by measuring connectivity amongst cortical regions implicated in semantic representation, then simulating healthy and disordered semantic processing in a deep neural network whose architecture mirrors this structure. The resulting neuro-computational model explains the full complement of neuroimaging and patient evidence adduced in support of both domain-specific and domain-general approaches, reconciling long-standing disputes about the nature and origins of this uniquely human cognitive faculty.

ERP signs of categorical and supra-categorical processing of visual information

BACKGROUND

The aim of the present study was to investigate to what extent shared and distinct brain mechanisms are possibly subserving the processing of visual supra-categorical and categorical knowledge as observed with event-related potentials of the brain. Access time to these knowledge types was also investigated. Picture pairs of animals, objects, and mixed types were presented. Participants were asked to decide whether each pair contained pictures belonging to the same category (either animals or man-made objects) or to different categories by pressing one of two buttons. Response accuracy and reaction times (RTs) were also recorded.

RESULTS

Both ERPs and RTs were grand-averaged separately for the same-different supra-categories and the animal-object categories. Behavioral performance was faster for more endomorphic pairs, i.e., animals vs. objects and same vs. different category pairs. For ERPs, a modulation of the earliest C1 and subsequent P1 responses to the same vs. different supra-category pairs, but not to the animal vs. object category pairs, was found. This finding supports the view that early afferent processing in the striate cortex can be boosted as a by-product of attention allocated to the processing of shapes and basic features that are mismatched, but not to their semantic quintessence, during same-different supra-categorical judgment. Most importantly, the fact that this processing accrual occurred independent of a traditional experimental condition requiring selective attention to a stimulus source out of the various sources addressed makes it conceivable that this processing accrual may arise from the attentional demand deriving from the alternate focusing of visual attention within and across stimulus categorical pairs' basic structural features. Additional posterior ERP reflections of the brain more prominently processing animal category and same-category pairs were observed at the N1 and N2 levels, respectively, as well as at a late positive complex level, overall most likely related to different stages of analysis of the greater endomorphy of these shape groups. Conversely, an enhanced fronto-central and fronto-lateral N2 as well as a centro-parietal N400 to man-made objects and different-category pairs were found, possibly indexing processing of these entities' lower endomorphy and isomorphy at the basic features and semantic levels, respectively.

CONCLUSION

Overall, the present ERP results revealed shared and distinct mechanisms of access to supra-categorical and categorical knowledge in the same way in which shared and distinct neural representations underlie the processing of diverse semantic categories. Additionally, they outlined the serial nature of categorical and supra-categorical representations, indicating the sequential steps of access to these separate knowledge types.

Disturbed interplay between mid- and high-level vision in ASD? Evidence from a contour identification task with everyday objects

Atypical visual processing in children with autism spectrum disorder (ASD) does not seem to reside in an isolated processing component, such as global or local processing. We therefore developed a paradigm that requires the interaction between different processes-an identification task with Gaborized object outlines-and applied this to two age groups of 6-to-10 and 10-to-14 year old children with and without ASD. Event history analyses demonstrated an identification disadvantage in the ASD group, which remained quite stable during the temporal unfolding of the outline. The typically developing group particularly outperformed the ASD group when more complex contours were shown. Together, our results suggest that the interplay between local and global processes and between bottom-up and top-down processes is disturbed in ASD.

The activation of alternative response candidates: when do doubts kick in?

In the current study, we investigated at which moment during visual object categorization alternative interpretations are most strongly activated. According to an early activation account, we are uncertain about how to interpret the visual information early in the categorization process. This uncertainty will vanish over time and therefore, the number of possible response candidates decreases over time. According to a late activation account, the visual information is categorized quickly, but after extensive viewing alternative interpretations become more strongly activated. Therefore, the number of possible response candidates increases over time. To increase perceptual uncertainty we used morphed figures composed of a dominant and nondominant object. The similarity rating between morphed figures and their nondominant object was taken as indicator for the activation of the nondominant response candidate: high similarity indicates that the nondominant object is relatively strongly activated as an alternative response candidate. Presentation times were varied in order to distinguish between the early and late activation account. Using a Bayesian model selection approach, we found support for the late activation account, but not for the early activation account. It thus seems that in a late stage of the categorization process the influence of the nondominant response candidate is strongest.

An event-related potential study of semantic style-match judgments of artistic furniture

This study investigates how semantic networks represent different artistic furniture. Event-related potentials (ERPs) were recorded while participants made style-match judgments for table and chair sets. All of the tables were in the Normal style, whereas the chairs were in the Normal, Minimal, ReadyMade, or Deconstruction styles. The Normal and Minimal chairs had the same rates of "match" responses, which were both higher than the rates for the ReadyMade and Deconstruction chairs. Compared with Normal chairs, the ERPs elicited by both ReadyMade chairs and Deconstruction chairs exhibited reliable N400 effects, which suggests that these two design styles were unlike the Normal design style. However, Minimal chairs evoked ERPs that were similar to the ERPs of Normal chairs. Furthermore, the N400 effects elicited by ReadyMade and Deconstruction chairs showed different scalp distributions. These findings reveal that semantic networks represent different design styles for items of the same category.

Categorization and action: what about object consistence?

Categorization studies have focused on the importance of a variety of perceptual properties (shape, size, weight). The present study explored whether the softness or hardness of an object might influence the way we categorize and consider category members. Of additional interest was whether information on consistence is automatically activated and whether it is modulated by the kind of task and of response modality. Three experiments demonstrated that information on consistence is automatically activated, and it helps us to distinguish between artefacts and natural objects. Interestingly, the results are in agreement with the simulation hypothesis; namely, when we consider artefacts, we simulate using them and information on their consistence is activated; this simulation is modulated by the task. The way we differently process artefacts and natural objects across the experiments confirms the simulation hypothesis and our sensitivity to the response modality.

Where is the semantic system? A critical review and meta-analysis of 120 functional neuroimaging studies

Semantic memory refers to knowledge about people, objects, actions, relations, self, and culture acquired through experience. The neural systems that store and retrieve this information have been studied for many years, but a consensus regarding their identity has not been reached. Using strict inclusion criteria, we analyzed 120 functional neuroimaging studies focusing on semantic processing. Reliable areas of activation in these studies were identified using the activation likelihood estimate (ALE) technique. These activations formed a distinct, left-lateralized network comprised of 7 regions: posterior inferior parietal lobe, middle temporal gyrus, fusiform and parahippocampal gyri, dorsomedial prefrontal cortex, inferior frontal gyrus, ventromedial prefrontal cortex, and posterior cingulate gyrus. Secondary analyses showed specific subregions of this network associated with knowledge of actions, manipulable artifacts, abstract concepts, and concrete concepts. The cortical regions involved in semantic processing can be grouped into 3 broad categories: posterior multimodal and heteromodal association cortex, heteromodal prefrontal cortex, and medial limbic regions. The expansion of these regions in the human relative to the nonhuman primate brain may explain uniquely human capacities to use language productively, plan, solve problems, and create cultural and technological artifacts, all of which depend on the fluid and efficient retrieval and manipulation of semantic knowledge.

Too little, too late: reduced visual span and speed characterize pure alexia

Whether normal word reading includes a stage of visual processing selectively dedicated to word or letter recognition is highly debated. Characterizing pure alexia, a seemingly selective disorder of reading, has been central to this debate. Two main theories claim either that 1) Pure alexia is caused by damage to a reading specific brain region in the left fusiform gyrus or 2) Pure alexia results from a general visual impairment that may particularly affect simultaneous processing of multiple items. We tested these competing theories in 4 patients with pure alexia using sensitive psychophysical measures and mathematical modeling. Recognition of single letters and digits in the central visual field was impaired in all patients. Visual apprehension span was also reduced for both letters and digits in all patients. The only cortical region lesioned across all 4 patients was the left fusiform gyrus, indicating that this region subserves a function broader than letter or word identification. We suggest that a seemingly pure disorder of reading can arise due to a general reduction of visual speed and span, and explain why this has a disproportionate impact on word reading while recognition of other visual stimuli are less obviously affected.

Category-specificity in visual object recognition

Are all categories of objects recognized in the same manner visually? Evidence from neuropsychology suggests they are not: some brain damaged patients are more impaired in recognizing natural objects than artefacts whereas others show the opposite impairment. Category-effects have also been demonstrated in neurologically intact subjects, but the findings are contradictory and there is no agreement as to why category-effects arise. This article presents a pre-semantic account of category-effects (PACE) in visual object recognition. PACE assumes two processing stages: shape configuration (the binding of shape elements into elaborate shape descriptions) and selection (among competing representations in visual long-term memory), which are held to be differentially affected by the structural similarity between objects. Drawing on evidence from clinical studies, experimental studies with neurologically intact subjects and functional imaging studies, it is argued that PACE can account for category-effects at both behavioural and neural levels in patients and neurologically intact subjects. The theory also accounts for the way in which category-effects are affected by different task parameters (the degree of perceptual differentiation called for), stimulus characteristics (whether stimuli are presented as silhouettes, full line-drawings, or fragmented forms), stimulus presentation (stimulus exposure duration and position) as well as interactions between these parameters.

Shared and/or separate representations of animate/inanimate categories: An ERP study

This paper presents an ERP experiment examining the underlying nature of semantic representation of animate and inanimate objects. Time-locking ERP signatures to the onset of visual stimuli we found topological similarities in animate and inanimate object processing. Moreover, when mapping more general to more specific representation (auditory to visual stimuli) we found no difference between animates and in animates in the N400 amplitude either. This study provides further evidence for the theory of unitary semantic organization, but no support for the feature-based prediction of segregated conceptual organization. Surprisingly, it was also found that the P600 component, which has been thus far mostly related to syntactic processing to be a sensitive index of conceptual processing. The most apparent difference regarding P600 component was found between animate and inanimate matches, whereby animate matches elicited more positive ERP signatures in comparison to inanimate matches.

Mid-fusiform activation during object discrimination reflects the process of differentiating structural descriptions

The present study explored constraints on mid-fusiform activation during object discrimination. In three experiments, participants performed a matching task on simple line configurations, nameable objects, three dimensional (3-D) shapes, and colors. Significant bilateral mid-fusiform activation emerged when participants matched objects and 3-D shapes, as compared to when they matched two-dimensional (2-D) line configurations and colors, indicating that the mid-fusiform is engaged more strongly for processing structural descriptions (e.g., comparing 3-D volumetric shape) than perceptual descriptions (e.g., comparing 2-D or color information). In two of the experiments, the same mid-fusiform regions were also modulated by the degree of structural similarity between stimuli, implicating a role for the mid-fusiform in fine differentiation of similar visual object representations. Importantly, however, this process of fine differentiation occurred at the level of structural, but not perceptual, descriptions. Moreover, mid-fusiform activity was more robust when participants matched shape compared to color information using the identical stimuli, indicating that activity in the mid-fusiform gyrus is not driven by specific stimulus properties, but rather by the process of distinguishing stimuli based on shape information. Taken together, these findings further clarify the nature of object processing in the mid-fusiform gyrus. This region is engaged specifically in structural differentiation, a critical component process of object recognition and categorization.

Identification of everyday objects on the basis of fragmented outline versions

Although Attneave (1954 Psychological Review 61 183 193) and Biederman (1987 Psychological Review 94 115-147) have argued that curved contour segments are most important in shape perception, Kennedy and Domander (1985 Perception 14 367-370) showed that fragmented object contours are better identifiable when straight segments are shown. We used the set of line drawings published by Snodgrass and Vanderwart (1980 Journal of Experimental Psychology: Human Learning and Memory 6 174-215), to make outline versions that could be used to investigate this issue with a larger and more heterogeneous stimulus set. Fragments were placed either around the 'salient' points or around the midpoints (points midway between two salient points), creating curved versus relatively straight fragments when the original outline was fragmented (experiment 1), or angular and straight fragments when straight-line versions were fragmented (experiment 2). We manipulated fragment length in each experiment except the last one, in which we presented only selected points (experiment 3). While fragmented versions were on average more identifiable when straight fragments were shown, certain objects were more identifiable when the curved segments or the angles were shown. A tentative explanation of these results is presented in terms of an advantage for straight segments during grouping processes for outlines with high part salience, and an advantage for curved segments during matching processes for outlines with low part salience.

A review of functional imaging studies on category specificity

A review of 20 functional imaging studies that compared visual processing of natural objects and artifacts in normal subjects is presented. The studies included fulfilled three criteria: (i) they used pictures as stimuli, (ii) they were based on direct contrasts between categories, and (iii) they reported findings in Talairach space. Not a single area is consistently activated for a given category across all studies. In contrast, 11 out of 29 regions are reported activated by both artifacts and natural objects. It is argued that the inconsistency is unlikely to reflect differences between studies in task requirements alone. Rather, the most likely causes of the inconsistency are: (i) adoption of liberal statistical thresholds that may yield false-positive activations, (ii) limited sensitivity due to few observations, and (iii) failure to match categories on confounding variables such as familiarity and visual complexity. Of the most consistent activations found, none appear to be selective for natural objects or artifacts. The findings reviewed are compatible with theories of category specificity that assume a widely distributed conceptual system not organized by category.

An attenuation of the 'normal' category effect in patients with Alzheimer's disease: a review and bootstrap analysis

There is a consensus that Alzheimer's disease (AD) impairs semantic information, with one of the first markers being anomia i.e. an impaired ability to name items. Doubts remain, however, about whether this naming impairment differentially affects items from the living and nonliving knowledge domains. Most studies have reported an impairment for naming living things (e.g. animals or plants), a minority have found an impairment for nonliving things (e.g. tools or vehicles), and some have found no category-specific effect. A survey of the literature reveals that this lack of agreement may reflect a failure to control for intrinsic variables (such as familiarity) and the problems associated with ceiling effects in the control data. Investigating picture naming in 32 AD patients and 34 elderly controls, we used bootstrap techniques to deal with the abnormal distributions in both groups. Our analyses revealed the previously reported impairment for naming living things in AD patients and that this persisted even when intrinsic variables were covaried; however, covarying control performance eliminated the significant category effect. Indeed, the within-group comparison of living and nonliving naming revealed a larger effect size for controls than patients. We conclude that the category effect in Alzheimer's disease is no larger than is expected in the healthy brain and may even represent a small diminution of the normal profile.

The impact of colour, spatial resolution, and presentation speed on category naming

Studies of neurological patients with category-specific agnosia have provided important contributions to our understanding of object recognition, although the meaning of such disorders is still hotly debated. One crucial line of research for our understanding of category effects, is through the examination of category biases in healthy normal subjects. This approach has, however, led to contradictory findings with advantages both for natural kinds and for man-made things being documented in healthy subjects. It has been proposed that task conditions may influence the direction of advantage (Gerlach, 2001) and in particular, that sub-optimal viewing conditions underpin natural kinds advantages, while man-made advantages emerge under more optimal viewing conditions. In two experiments with normal subjects, we examined the roles played by spatial resolution (blurring), stimulus type (colour and texture), and speed of presentation in picture naming across category. In both experiments, healthy subjects showed a natural kind advantage for original stimuli and for blurred colour stimuli (at slow and fast presentation speeds), while an advantage for man-made things emerged for line-drawings that were blurred and presented slowly. The implications for category-specific object recognition deficits are discussed.

Shape configuration and category-specificity

We examined the neural correlates of visual shape configuration, the binding of local shape characteristics into wholistic object descriptions, by comparing the regional cerebral blood flow associated with recognition of outline drawings and fragmented drawings. We found no areas that responded more to fragmented drawings than to outline drawings even though fragmentation had a clear impact on recognition performance. Instead, a region extending from the inferior occipital gyri to the middle parts of the fusiform gyri was activated during shape configuration of both outline drawings and fragmented drawings. We also examined whether fragmentation had different impact on the recognition of natural objects and artefacts and found that recognition of artefacts was more affected by fragmentation than recognition of natural objects. Thus, the usual finding of an advantage for artefacts in difficult object decision tasks, which is also found in the present experiments with outlines, is reversed when the stimuli are fragmented. This interaction between category (natural versus artefacts) and stimulus type (outlines versus fragmented forms) is in accordance with predictions derived from a recent account of category-specificity and lends support to the notion that category-specific impairments can occur for both natural objects and artefacts following damage to pre-semantic stages in visual object recognition. The implications of the present findings are discussed in relation to theories of perceptual organization, visual object recognition and category-specificity.