Finally, we have found that brain regions traditionally thought to be involved in explicit memory play a role in object detection. In this chapter, our work examining how past experience and meaning affects object detection is reviewed and integrated with current research from other laboratories.
My goal is for this review to serve as a springboard for research aimed at a deeper understanding of the dynamical interactions involved in object perception.
Past experience and meaning affect object detection : A hierarchical Bayesian approach. N2 - For human perceivers, object perception seems immediate and unambiguous.
AB - For human perceivers, object perception seems immediate and unambiguous. Past experience and meaning affect object detection: A hierarchical Bayesian approach. Cognitive Science. Participants were shown gray line drawings of 60 common objects containing what the researchers defined as distinct, nameable, spatially discrete features. During the experimental task, EEG data were collected while participants were presented with a line drawing stimulus for ms. Following a ms pause, participants were cued to generate a mental image of the previously viewed line drawing according to the displayed holistic or partial-image word cue.
These lexical cues referenced either the canonical name of an entire object or the name of one of its semantically meaningful parts e. A button press on a response box indicated the starting time point for each imaginary episode.
The results of the study revealed that the greatest differences between perceptual and imaginary tasks existed within the theta and alpha band spectra. The authors speculate that the early time onset associated with the partial imagery condition indicates that partial visual information manifests independently from holistic imagery generation, and that stronger theta power in this condition reflects more complex processing required to retrieve object details.
These findings suggest that, despite its apparent independence from holistic information, partial imagery formation appears to involve complex interactions with its associated holistic context.
This could have important implications for understanding the relationship between distinct diagnostic features and the whole representation to which they are related. Similar to Ullman et al. This is an important consideration when attempting to identify minimally diagnostic regions, but Li et al. The word cue approach employed to manipulate the experimental mental imagery condition in Li et al. The use of word cues based on nameable object parts forces an artificial separation of an image into spatially discrete features defined by recognizable, but arbitrary, structural features, the interactive relationship between each and the overall holistic representation is unclear.
In addition, a computational model used to decode human brain activity to predict fMRI activation associated with the meaning of nouns Mitchell et al. Therefore, although the object part manipulation applied in Li et al. Functional near-infrared spectroscopy fNIRS is a relatively new technology that is rapidly gaining popularity due to its portability and flexibility of experimental application.
This system combines the spatial sensitivity of fMRI with the convenience and temporal resolution of EEG through non-invasive measurement of the diffusion rate of near-infrared light as it is projected through the skull.
Recordings collected from fNIRS are influenced by the relative concentration of oxygenated and deoxygenated hemoglobin in cortical blood flow, and are thus conceptualized as an indirect measurement of neural activity Kamran and Hong, This comparatively recent methodology has yet to be directly applied to tasks involving visual mental imagery generation.
However, studies of infant perception, as well as a substantial base of motor imagery decoding literature, show mixed evidence for hemodynamic responses recorded via fNIRS as an index of private mental and visual processes. A study of visual processing in infants employed fNIRS in an investigation of the neural correlates underlying object individuation Wilcox et al. Using a version of the narrow-screen task Wilcox and Baillargeon, , infants were familiarized to two featurally distinct objects, a ball and a box, as they successively appeared from opposite sides of a narrow or wide screen.
Behavioral response times revealed that infants looked longer at the narrow screen condition, suggesting that the infants were able to discriminate the stimuli as two distinct objects which could not logically fit behind the screen along the same axis simultaneously.
Variations in hemodynamic response as measured by NIRS during narrow screen trials were localized to the primary visual and inferior temporal cortex, indicating that object individuation is associated with unique, detectable patterns of neural activity in these regions Wilcox et al.
Although the researchers acknowledge that more work remains to be done before the connection between variations in oxygenated and deoxygenated hemoglobin and cognition are well-understood but see Chen et al. Furthermore, the capacity for fNIRS to discriminate amongst neural activity influenced by variations in distinct local features shows promise for detecting distinguishing features in object recognition.
Several concerns remain when considering fNIRS as a measure of mental imagery. On the one hand, a relatively established body of work investigating motor imagery may hold clues for guiding future application of this technology to object representations. The bulk of this literature focuses on decoding motor imagery for application to brain-computer interface technology see Naseer and Hong, , for review. Accordingly, NIRS measurements are frequently recorded from regions of motor cortex, which tend to be easily penetrated by near-infrared.
Although experiments that apply fNIRS to motor imagery decoding do not attempt to access visual mental imagery directly, their results demonstrate the potential for fNIRS data to facilitate a reliable decoding of private internal events. This may indicate that the same methods could be applied to visual object mental imagery, so long as relevant cortical surfaces can be reached to be discussed later in this section.
If the neural substrates of mental imagery can indeed be accessed, the high temporal and spatial resolution afforded by fNIRS technology may be a worthwhile avenue through which to pursue diagnostic features represented in mental imagery. Unfortunately, skepticism on the accuracy and utility of fNIRS recordings, even in motor imagery, remains e. The investigation of mental imagery through brain activity is clearly advantageous in that these methods do not require individuals to explicitly communicate their private mental experiences.
Despite the demonstrated success of neurophysiological recording methods in accessing mental imagery, there are still several limitations that must be considered when examining their implications for identifying distinguishing features in object representations. For example, similar to the stimulus set utilized by Behroozi and Daliri , successful decoding of category or identity information in fMRI often relies on a collection of images from which to decode responses, with few exceptions e.
This requirement restricts the flexibility with which neural imaging methods can index natural imagery generation in real-world scenarios, which contain multiple variables and a huge set of possibilities for visual stimuli. However, it should be noted that some studies have managed to push the boundaries of this set to impressive limits and still report success in decoding components of mental imagery e.
In addition to fMRI, EEG dynamics are likely capable of capturing effects caused by the processing of spatially defined object components Li et al. However, as mentioned above, manipulating the spatial resolution of defined features via verbal cues poses several challenges, and, as of yet, EEG has demonstrated success mostly at the level of general or superordinate classification.
The issue of avoiding artificial biases introduced by arbitrary delineation of structural features extends to any perceptual assessment of mental imagery. Careful and creative experimental designs are required in order to achieve a method capable of accessing discrete diagnostic visual features as they occur naturally and at a level accessible by EEG.
Although results from fNIRS studies show limited ability to categorize motor imagery and visual perceptual processes, there are several concerns that affect this area of research as applied to feature detection in object mental imagery. First, the nature of infrared light penetration employed by fNIRS restricts recordings to areas of cortex lying closely beneath the skull approximately 2—3 cm below the cortical surface Wilcox et al.
Fortunately, there is evidence that cortical visual processing areas are accessible via near-infrared. In addition to recordings obtained from primary and ventral visual cortex in infants Meek et al. Furthermore, fNIRS recordings collected from adult primary visual cortex and supplementary neurophysiological measures show that fNIRS is capable of revealing patterns of stimulus selectivity as well as area specificity Chen et al. However, considering the multiple findings which implicate areas outside of early visual cortex as significant contributors to visual mental images see Vetter et al.
An additional advantage of neuroimaging methods is that their data lend themselves to a wide range of statistical analyses which allow interpretations of complex activation patterns to identify correlations between visual and semantic information.
Many of these techniques, including MVPA and support vector machine learning, allow multiple factors to be considered when correlating brain activity with category information see De Martino et al.
The variety of statistical methods which may be applied to neuroimaging data enhances their potential for drawing inferences between neural activity and semantic object recognition processes, which may eventually allow specific predictions to be made regarding object feature information within mental images. Overall, fMRI neural recordings of perceptual and mental visual processing hold great potential for indexing distinguishing features in object representations, whereas EEG and fNIRS appear to be weaker methods.
The sensitivity to object content demonstrated by fMRI data clearly suggests their utility for accessing the discrete visual elements necessary for distinguishing amongst different object categories.
Results from EEG reflect partial imagery effects in brain wave dynamics, but current work is limited to general and disparate effects in holistic vs. In addition, there is reason to believe that fNIRS technology may be capable of recording mental imagery as expressed within adult primary visual cortical regions. However, more work remains to be done before fNIRS can be confidently applied to imaginary representations in either the motor or object imagery domains.
The aim of this review is twofold: to propose that mental imagery is an advantageous and valid method for assessing diagnostic object features, and to demonstrate that, in spite of the present lack of direct investigations of diagnostic features within mental imagery, evidence of their relationship and the tools best suited to their study are suggested by existing literature.
Each measurement method offers its own unique advantages and weaknesses for exploring the role of diagnostic visual components in object processing see Table 1. The early, underdeveloped state of this field favors a systematic methodological approach capable of capturing a broad range of information.
In order to achieve this, measurement approaches should be combined with the aim of capitalizing on methodological strengths and compensating for weaknesses, with an emphasis for collecting large and varied amounts of data. Here, overall implications and the utility of each method for exploring distinguishing features in visual imagery are summarized and suggestions are made for future directions.
TABLE 1. Methodological pros and cons for accessing object mental imagery content. The fundamental methodological issues associated with asking individuals to recreate or verbalize their internal experiences limit the possibilities for applying this tool as a direct individual assessment of diagnostic features.
Pooling responses over a large group provides the best opportunity to reveal meaningful trends regarding the nature of classificatory object features. In addition, the ease of use and rapid administration of self-report instruments facilitate their combination with other forms of measurement. The consideration of questionnaire responses alongside neurophysiological data collected from methods such as eye tracking, EEG, and fMRI provides several benefits:.
Studies of motor movement behavior implicate a significant relationship between goal-directed actions and mental representations that is mediated by the functional end of manipulable objects. When acquired without physical practice, however, this relationship appears to be limited to large-scale, goal-directed behaviors that are expected to produce an appropriate consequence.
Because precise motor behaviors such as grip have not been shown to share a direct correlation with imagery-acquired mental representations, this approach may only appropriate for items that are associated with obvious and unique goal-directed movements, such as tools. Nevertheless, the observed correlation between spatially discrete, structural object components and motor behavior supports the likelihood that distinct features exist in mental representations, and that cognitively significant features may be indirectly expressed through physical actions.
Eye tracking findings indicate that saccades are similar across perceptual viewing and mental imagery states, and thus reflect meaningful cognitive processes. Although saccadic eye movements in either condition may implicate attention to spatial location more so than individual features, this relationship may be exploited for the benefit of diagnostic feature research.
If object features were to be reliably equated with distinct spatial locations in a manner that does not artificially modify the holistic representation of an object, saccadic eye movements may allow a more direct index of attention to salient object features rather than spatial location, which can then be tested for classificatory utility.
In addition, such a method may pre-empt possible confounds of covert attention by requiring overt eye movements in order to fixate discrete features.
This approach could be applied to perceptual object recognition tasks, the results of which could guide and be compared to analogous mental imagery investigations. A straightforward assessment of diagnostic features across perceptual and imagery conditions such as this may lead to further insight into the broader interactions between object features, perceptual processing, and mental imagery. Neural activity recordings obtained through fMRI constitute the most established area of research regarding internal visual images.
This method provides a direct index of mental imagery content while avoiding the complications associated with verbal or visual translation of private mental experiences. Category classifications achieved via blood-oxygen-level dependent fluctuations as measured by fMRI strongly implicate the presence of diagnostic feature information in mental images, as well as their expression through neural activity.
Distinct patterns of activity associated with particular categories of imagined objects have been identified and found to remain stable across individuals. Overall, fMRI techniques show notable promise for advancing the understanding of the role of diagnostic object features in object representations via the underlying neural mechanisms of mental imagery.
Preliminary findings from EEG studies also suggest that holistic and non-holistic imagery of partial object components may be reflected in brain wave dynamics, but specific feature information has yet to be identified, suggesting it may be best combined with methods such as fMRI. The detailed level of specificity accomplished through fMRI is limited by its lack of temporal resolution, whereas EEG is limited by spatial resolution.
Perceptual work conducted with fNIRS provides inconclusive evidence regarding hemodynamic responses as a reliable indicator of imagery-related activity. Although perceptual research employing fNIRS is still in its early stages, studies of object individuation and motor imagery suggest potential for future application of fNIRS to mental imagery events. Unfortunately, near-infrared is currently limited to the shallow cortical regions involved in visual processing, and the reliability of hemodynamic response for neural decoding remains to be validated.
In order to accommodate the widely distributed and varied neural correlates of mental imagery formation, fNIRS should be combined with broader measurement instruments such as EEG or fMRI in order to aid effective indexing of visualized object features. The methods and findings reviewed in this article are intended to support the feasibility and value of exploring perceptual diagnostic features through visual mental representations.
Internal visual experiences, which occur in the absence of perceptual input, have the potential to be uniquely informative for understanding the manner in which visual information is transformed and translated to produce semantically meaningful object representations. By taking advantage of the natural information filtration processes imposed by the physical and cognitive limitations of human visual and neuronal systems, it is possible to access a great deal of semantic information via a condensed, concentrated source, in the form of distinguishing features.
Mental imagery offers significant benefits over direct perceptual evaluations in that, during natural perceptual viewing, the potential amount of identifying object information is far greater than when an object is recalled through imagery alone, in part because irrelevant information has been discarded — or, at least, deemphasized — in mental representations.
By studying object recognition solely as it occurs during perceptual processes, the researcher is forced to consider an inordinate amount of possibilities when determining the properties of a stimulus most essential to its identification. However, this procedure poorly reflects natural visual processes. By approaching visual perception from the level of its ultimate goal — the mental representation — and evaluating the relationship between the initial input and its final output, the investigator may reach a more complete and accurate understanding of the interaction between vision and cognition.
SR developed the concept, conducted the research, and wrote the text of this manuscript. The author declares that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
The reviewer DL and handling Editor declared their shared affiliation, and the handling Editor states that the process nevertheless met the standards of a fair and objective review. The author would like to thank Dr. Anthony Cate for his guidance during the planning and preparation of this work, as well as Drs.
Diana, Bell, and LaConte for their help in the development of concepts and in finding relevant research. Albers, A. Shared representations for working memory and mental imagery in early visual cortex. Anderson, J. A spreading activation theory of memory.
Verbal Learn. Verbal Behav. Andrade, J. Ballesteros, S. Selective attention affects conceptual object priming and recognition: a study with young and older adults. Barrett, T. Infant Behav. Bartolomeo, P. The neural correlates of visual mental imagery: an ongoing debate. Cortex 44, — Baruch, O. Attention to distinguishing features in object recognition. Becker, M. Object-intrinsic oddities draw early saccades.
Behrmann, M. Intact visual imagery and impaired visual perception in a patient with visual agnosia. Behroozi, M. Predicting brain states associated with object categories from fMRI data. Biederman, I. Recognition-by-components: a theory of human image understanding. Blajenkova, O. Object-spatial imagery: a new self-report imagery questionnaire.
Borst, G. Representations in mental imagery and working memory: evidence from different types of visual masks. Visual mental imagery and visual perception: structural equivalence revealed by scanning processes. The influence of color information on the recognition of color diagnostic and noncolor diagnostic objects. The role of color information on object recognition: a review and meta-analysis. Acta Psychol.
Brandt, S. Email: isabel. Barton J. Hanif H. Ashraf S. Relating visual to verbal semantic knowledge: The evaluation of object recognition in prosopagnosia. Brain , 12 , —, doi: Bilalic M. Langner R. Ulrich R.
Grodd W. Many faces of expertise: Fusiform face area in chess experts and novices. Journal of Neuroscience , 31 28 , —, doi: Bowles D. McKone E. Dawel A. Duchaine B. Palermo R. Schmalzl L. Cognitive Neuropsychology , 26 5 , —, doi: Brown A. Jensen J. Nikolova Y. Djurovic S. Agartz I. Server A. Genetic variants affecting the neural processing of human facial expressions: Evidence using a genome-wide functional imaging approach.
Translational Psychiatry , 2 7 , e, doi: Cho S. Wilmer J. Herzmann G. McGugin R. Fiset D. Van Gulick A. Colquitt J. Simmering M. Conscientiousness, goal orientation, and motivation to learn during the learning process: A longitudinal study. Journal of Applied Psychology , 83 4 , — Curby K. Glazek K. Gauthier I. A visual short-term memory advantage for objects of expertise.
Davis J. Dennett H. O'Connor K. O'Kearney R. Individual differences in the ability to recognise facial identity are associated with social anxiety. Edwards M. Susilo T.
Face aftereffects predict individual differences in face recognition ability. Psychological Science , 23 11 , —, doi: Tavashmi R. Hall A. Pidcock M. The Cambridge Car Memory Test: A task matched in format to the Cambridge Face Memory Test, with norms, reliability, sex differences, dissociations from face memory, and expertise effects. Behavior Research Methods , 44 2 , —, doi: Nakayama K. The Cambridge Face Memory Test: Results for neurologically intact individuals and an investigation of its validity using inverted face stimuli and prosopagnosic subjects.
Neuropsychologia , 44 4 , —, doi: Engle R. Tuholski S. Laughlin J. Conway A. Working memory, short-term memory, and general fluid intelligence: A latent-variable approach.
Journal of Experimental Psychology: General , 3 , — Ericsson K. Lehmann A. Expert and exceptional performance: Evidence of maximal adaptation to task constraints. Annual Review of Psychology , 47 , — A perceptual traffic-jam on Highway N Interference between face and car expertise.
Current Directions in Psychological Science , 14 1 , 30— Skudlarski P. Epstein R. Individual differences in FFA activity suggest independent processing at different spatial scales.
Cognitive, Affective, and Behavioral Neuroscience , 5 2 , — Curran T. Collins D. Perceptual interference supports a non-modular account of face processing. Nature Neuroscience , 6 4 , —, doi: Gore J. Anderson A. Expertise for cars and birds recruits brain areas involved in face recognition.
Nature Neuroscience , 3 2 , —, doi: Tarr M. Unraveling mechanisms for expert object recognition: Bridging brain activity and behavior. Germine L. Where cognitive development and aging meet: Face learning ability peaks after age Cognition , 2 , —, doi: Harel A. Gilaie-Dotan S. Malach R. Bentin S. Top-down engagement modulates the neural expressions of visual expertise. Cerebral Cortex , 20 10 , —, doi: Harley E. Pope W. Villablanca J. Mumford J. Suh R. Mazziotta J. Engagement of fusiform cortex and disengagement of lateral occipital cortex in the acquisition of radiological expertise.
Philosophers of perception frequently assume that we see actual states of affairs, or facts. Call this claim factualism. In his book, William Fish suggests that factualism is supported by … Expand. A behavioral task sets an upper bound on the time required to access object memories before object segregation. Neurophysiological evidence for the influence of past experience on figure-ground perception. Toward a Dynamical View of Object Perception. View 3 excerpts, references results. Language can boost otherwise unseen objects into visual awareness.
The Ground Side of an Object. Neural evidence for competition-mediated suppression in the perception of a single object.
0コメント