The Experts below are selected from a list of 306 Experts worldwide ranked by ideXlab platform
Takashi Yamauchi - One of the best experts on this subject based on the ideXlab platform.
-
Category labels versus feature labels: Category labels polarize inferential predictions.
Memory & cognition, 2008Co-Authors: Takashi YamauchiAbstract:What makes Category labels different from feature labels in predictive inference? This study suggests that Category labels tend to make inductive reasoning polarized and homogeneous. In two experiments, participants were shown two schematic pictures of insects side by side and predicted the value of a hidden feature of one insect on the basis of the other insect. Arbitrary verbal labels were shown above the two pictures, and the meanings of the labels were manipulated in the instructions. In one condition, the labels represented the Category Membership of the insects, and in the other conditions, the same labels represented attributes of the insects. When the labels represented Category Membership, participants’ responses became substantially polarized and homogeneous, indicating that the mere reference to Category Membership can modify reasoning processes.
-
Tracking mouse movement in feature inference: Category labels are different from feature labels
Memory & Cognition, 2007Co-Authors: Takashi Yamauchi, Nicholas KohnAbstract:In this article, we examine the role of Category labels in inductive inference. Some leading research has suggested that information about Category Membership works just like any other feature in categorical inductions, whereas other research has proposed that the influence of Category Membership on induction goes beyond that of other features. To investigate these claims further, we developed an online measure of judgments that is akin to eyetracking. The judgment results and the mouse-tracking data jointly support the view that Category labels do affect inductive inferences in a way distinct from that for feature information. When arbitrary labels conveyed Category Membership information, participants viewed these labels more often and earlier in a trial, in comparison with cases in which the same labels conveyed non-Membership information. Our results suggest that Category Membership information works like a guide for inference. An ecological rationale for this induction strategy is also discussed.
-
Categories and Feature Inference: Category Membership and a Reasoning Bias
2005Co-Authors: Takashi YamauchiAbstract:Categories and Feature Inference: Category Membership and a Reasoning Bias Takashi Yamauchi (tya@psyc.tamu.edu) Na-Yung Yu (dbskdud40@tamu.edu) Department of Psychology, Mail Stop 4235 Texas AM Osherson, et al., 1990; Sloman, 1993, 1998; Sloutsky, 2003). The information about Category Membership may draw more attention; yet, people interpret Category Membership as they do for other perceptual and conceptual attributes. In this paper, we propose that Category Membership plays a special role in feature inferences, and molds people’s inferential behavior in a way other regular attributes cannot. Specifically, we propose that the awareness of Category Membership generally creates a reasoning strategy and biases people’s inferential behavior. Consider a simple prediction task in which one infers the value of an unknown feature on the basis of another stimulus (Figure 1) (Murphy & Ross, 1994; Yamauchi & Markman, 2000). In one case, two stimuli have the same arbitrary label “monek” (Figure 1a); in the other case, two stimuli have different labels “moneke” and “plaple.” In this circumstance, we think that people generally apply the following reasoning rules proportional to the extent to which the two labels carry the information about Category Membership: Rule 1 – if two items belong to the same Category, then the two items have characteristics in common; Rule 2 – if two items belong to different categories, then the two items have different characteristics. Undoubtedly, this reasoning strategy is erroneous because the members of natural categories are organized probabilistically, and the shared label does not necessarily guarantee shared features (and vice versa) (Rosch & Mervis, 1975). Moreover, psychological responses that we can observe in empirical studies are characteristically probabilistic. Therefore, such extreme “Category-based” responses would rarely happen. However, we think that there is a cognitive bias to apply this “reasoning habit” in feature inferences when Category information is transparent. This reasoning strategy reflects the mutual-exclusivity constraint suggested by E. Markman (1989) and the psychological essentialism assumption suggested by Medin and Otorny (1989), and Gelman (2003). For example, the reason why shared labels lead to shared features is because a Category is bound by some unknown or unknowable essential features, and these essential features generate other features. Likewise, two categories are viewed as mutually exclusive because they are bound by two sets of essentially different features. We hypothesize that the mere presence of Category labels promotes this rule-based reasoning strategy and generates polarity and uniformity in feature inference (e.g., Goldstone, 1994; Tajfel, 1963). For example, by applying this induction strategy, people accentuate the difference between two groups (i.e., polarity hypothesis), and discount perceptual variability of individual stimuli (i.e., uniformity hypothesis). We tested whether or not such reasoning biases would appear when the arbitrary labels carry Category Membership information. Experiment In our experiment, participants received pairs of a sample stimulus and a test stimulus one pair at a time (Figure 1), and predicted the feature value of a test stimulus on the basis of the sample stimulus. The stimuli were schematic illustrations of cartoon bugs, which were composed of 5 feature dimensions with binary values (Table 1). Twenty test stimuli were presented twice. In one case, a test stimulus was paired with a sample stimulus that had the same label, and in the other case, the same test stimulus was paired with a sample stimulus that had a different label (Figures 1a & 1b).
-
Inference Using Categories
Journal of experimental psychology. Learning memory and cognition, 2000Co-Authors: Takashi Yamauchi, Arthur B. MarkmanAbstract:How do people use Category Membership and similarity for making inductive inferences? The authors addressed this question by examining the impact of Category labels and Category features on inference and classification tasks that were designed to be comparable. In the inference task, participants predicted the value of a missing feature of an item given its Category label and other feature values. In the classification task, participants predicted the Category label of an item given its feature values. The results from 4 experiments suggest that Category Membership influences inference even when similarity information contradicts the Category label. This tendency was stronger when the Category label conveyed class inclusion information than when the label reflected a feature of the Category. These findings suggest that Category Membership affects inference beyond similarity and that Category labels and Category features are 2 different things.
Todd M Gureckis - One of the best experts on this subject based on the ideXlab platform.
-
sparse Category labels obstruct generalization of Category Membership
Cognitive Science, 2012Co-Authors: John V Mcdonnell, Carol A Jew, Todd M GureckisAbstract:Sparse Category labels obstruct generalization of Category Membership John V. McDonnell (john.mcdonnell@nyu.edu) Carol A. Jew (carol.jew@nyu.edu) Todd M. Gureckis (todd.gureckis@nyu.edu) New York University, Department of Psychology 6 Washington Place, New York, NY 10003 USA Abstract child’s pre-linguistic grouping of objects in their environment into classes. A similar position is advocated by a number of in uen- tial theories of Category learning which hold that supervised and unsupervised learning are subserved by a single under- lying learning process (e.g., the rational model of categoriza- tion, Anderson, 1991; or the Supervised and Unsupervised STrati ed Adaptive Incremental Network, abbreviated - , Love, Medin, & Gureckis, 2004). Such models naturally predict that semi-supervised learning should not only be pos- sible, but may be the primary way in which people learn cate- gories and their respective names. Studies of human Category learning typically focus on situa- tions where explicit Category labels accompany each example (supervised learning) or on situations were people must infer Category structure entirely from the distribution of unlabeled examples (unsupervised learning). However, real-world cate- gory learning likely involves a mixture of both types of learning (semi-supervised learning). Surprisingly, a number of recent ndings suggest that people have difficulty learning in semi- supervised tasks. To further explore this issue, we devised a Category learning task in which the distribution of labeled and unlabeled items suggested alternative organizations of a cate- gory. is design allowed us to determine whether learners combined information from both types of episodes via their patterns of generalization at test. In contrast with the prediction of many models, we nd little evidence that unlabeled items in- uenced categorization behavior when labeled items were also present. Keywords: Semi-supervised Category learning; rule induction; unsupervised learning Can people acquire categories via semi-supervised learning? Despite these arguments, recent empirical attempts to demon- strate semi-supervised Category learning in the lab have met with mixed success. For example, Vandist, De Schryver, and Rosseel (2009) found that adding unlabeled training exam- ples to a mostly supervised task offered no additional bene- t beyond learning from only the supervised trials. However, the Category structures they tested (known as Information- Integration categories) are typically difficult for people to learn even in fully unsupervised settings (Ashby, Queller, & Berretty, 1999), which may explain the limited impact that the unlabeled examples had. On the other hand, Kalish, Rogers, Lang, and Zhu (2011) showed that aer learning a simple Category distinction on a single dimension from a small set of labeled examples, par- ticipants’ estimate of the Category boundary could be shied by the presentation of a large number of unlabeled examples whose distribution was shied compared to the labeled set (see also Lake & McClelland, 2011). While this study provides some evidence of semi-supervised learning, there remain al- ternative explanations of the effect. For example, since the central tendency of both categories are shied in these studies it is unclear whether people are separately updating each cat- egory representation or responding to the global shi in the stimulus space. Finally, Rogers, Kalish, Gibson, Harrison, and Zhu (2010) compared learning in a semi-supervised learning condition with a fully supervised condition. In this study, adding un- labeled items to a supervised Category learning task caused faster learning only when trials were speeded. However, the question of whether people can integrate labeled and unla- beled training examples is logically separate from claims about Introduction Category learning is a critical cognitive ability which is cen- tral to many aspects of cognition. As a result, considerable re- search over the last – years has explored the psychology of Category learning using laboratory tasks. e majority of this work can be divided into two groups. Most research has fo- cused on supervised learning tasks where corrective feedback or Category labels are presented following or alongside each observation of a stimulus (e.g., Medin & Schaffer, 1978; Nosof- sky, 1986). More recently, there has been an interest in unsu- pervised learning, wherein participants must organize exam- ples in the absence of explicit instruction using the distribu- tional properties of the stimuli (e.g., Clapper & Bower, 1994; Love, 2002; Pothos et al., 2011). However, neither of these sit- uations adequately re ect the problem of real world Category learning, in which feedback is not altogether absent nor always present, but is typically sparse and intermittent. Such tasks re- quire learners to combine information from both labeled and unlabeled episodes. In machine learning, this problem is fre- quently studied under the name semi-supervised learning (for review, see Zhu, 2005). Aside from offering a more ecologically relevant approach to the study of Category learning, the study of semi-supervised learning has important implications for theories of human concept learning. Consider the problem of learning a con- crete noun such as horse. One proposal is that word learn- ing essentially links sound tokens (words) to already-acquired hypotheses or representations (Bloom, 2000; Gentner, 1982). Under this view, the label information from a teacher or par- ent about a single example horse must be integrated with the
-
CogSci - Sparse Category labels obstruct generalization of Category Membership
Cognitive Science, 2012Co-Authors: John V Mcdonnell, Carol A Jew, Todd M GureckisAbstract:Sparse Category labels obstruct generalization of Category Membership John V. McDonnell (john.mcdonnell@nyu.edu) Carol A. Jew (carol.jew@nyu.edu) Todd M. Gureckis (todd.gureckis@nyu.edu) New York University, Department of Psychology 6 Washington Place, New York, NY 10003 USA Abstract child’s pre-linguistic grouping of objects in their environment into classes. A similar position is advocated by a number of in uen- tial theories of Category learning which hold that supervised and unsupervised learning are subserved by a single under- lying learning process (e.g., the rational model of categoriza- tion, Anderson, 1991; or the Supervised and Unsupervised STrati ed Adaptive Incremental Network, abbreviated - , Love, Medin, & Gureckis, 2004). Such models naturally predict that semi-supervised learning should not only be pos- sible, but may be the primary way in which people learn cate- gories and their respective names. Studies of human Category learning typically focus on situa- tions where explicit Category labels accompany each example (supervised learning) or on situations were people must infer Category structure entirely from the distribution of unlabeled examples (unsupervised learning). However, real-world cate- gory learning likely involves a mixture of both types of learning (semi-supervised learning). Surprisingly, a number of recent ndings suggest that people have difficulty learning in semi- supervised tasks. To further explore this issue, we devised a Category learning task in which the distribution of labeled and unlabeled items suggested alternative organizations of a cate- gory. is design allowed us to determine whether learners combined information from both types of episodes via their patterns of generalization at test. In contrast with the prediction of many models, we nd little evidence that unlabeled items in- uenced categorization behavior when labeled items were also present. Keywords: Semi-supervised Category learning; rule induction; unsupervised learning Can people acquire categories via semi-supervised learning? Despite these arguments, recent empirical attempts to demon- strate semi-supervised Category learning in the lab have met with mixed success. For example, Vandist, De Schryver, and Rosseel (2009) found that adding unlabeled training exam- ples to a mostly supervised task offered no additional bene- t beyond learning from only the supervised trials. However, the Category structures they tested (known as Information- Integration categories) are typically difficult for people to learn even in fully unsupervised settings (Ashby, Queller, & Berretty, 1999), which may explain the limited impact that the unlabeled examples had. On the other hand, Kalish, Rogers, Lang, and Zhu (2011) showed that aer learning a simple Category distinction on a single dimension from a small set of labeled examples, par- ticipants’ estimate of the Category boundary could be shied by the presentation of a large number of unlabeled examples whose distribution was shied compared to the labeled set (see also Lake & McClelland, 2011). While this study provides some evidence of semi-supervised learning, there remain al- ternative explanations of the effect. For example, since the central tendency of both categories are shied in these studies it is unclear whether people are separately updating each cat- egory representation or responding to the global shi in the stimulus space. Finally, Rogers, Kalish, Gibson, Harrison, and Zhu (2010) compared learning in a semi-supervised learning condition with a fully supervised condition. In this study, adding un- labeled items to a supervised Category learning task caused faster learning only when trials were speeded. However, the question of whether people can integrate labeled and unla- beled training examples is logically separate from claims about Introduction Category learning is a critical cognitive ability which is cen- tral to many aspects of cognition. As a result, considerable re- search over the last – years has explored the psychology of Category learning using laboratory tasks. e majority of this work can be divided into two groups. Most research has fo- cused on supervised learning tasks where corrective feedback or Category labels are presented following or alongside each observation of a stimulus (e.g., Medin & Schaffer, 1978; Nosof- sky, 1986). More recently, there has been an interest in unsu- pervised learning, wherein participants must organize exam- ples in the absence of explicit instruction using the distribu- tional properties of the stimuli (e.g., Clapper & Bower, 1994; Love, 2002; Pothos et al., 2011). However, neither of these sit- uations adequately re ect the problem of real world Category learning, in which feedback is not altogether absent nor always present, but is typically sparse and intermittent. Such tasks re- quire learners to combine information from both labeled and unlabeled episodes. In machine learning, this problem is fre- quently studied under the name semi-supervised learning (for review, see Zhu, 2005). Aside from offering a more ecologically relevant approach to the study of Category learning, the study of semi-supervised learning has important implications for theories of human concept learning. Consider the problem of learning a con- crete noun such as horse. One proposal is that word learn- ing essentially links sound tokens (words) to already-acquired hypotheses or representations (Bloom, 2000; Gentner, 1982). Under this view, the label information from a teacher or par- ent about a single example horse must be integrated with the
John V Mcdonnell - One of the best experts on this subject based on the ideXlab platform.
-
sparse Category labels obstruct generalization of Category Membership
Cognitive Science, 2012Co-Authors: John V Mcdonnell, Carol A Jew, Todd M GureckisAbstract:Sparse Category labels obstruct generalization of Category Membership John V. McDonnell (john.mcdonnell@nyu.edu) Carol A. Jew (carol.jew@nyu.edu) Todd M. Gureckis (todd.gureckis@nyu.edu) New York University, Department of Psychology 6 Washington Place, New York, NY 10003 USA Abstract child’s pre-linguistic grouping of objects in their environment into classes. A similar position is advocated by a number of in uen- tial theories of Category learning which hold that supervised and unsupervised learning are subserved by a single under- lying learning process (e.g., the rational model of categoriza- tion, Anderson, 1991; or the Supervised and Unsupervised STrati ed Adaptive Incremental Network, abbreviated - , Love, Medin, & Gureckis, 2004). Such models naturally predict that semi-supervised learning should not only be pos- sible, but may be the primary way in which people learn cate- gories and their respective names. Studies of human Category learning typically focus on situa- tions where explicit Category labels accompany each example (supervised learning) or on situations were people must infer Category structure entirely from the distribution of unlabeled examples (unsupervised learning). However, real-world cate- gory learning likely involves a mixture of both types of learning (semi-supervised learning). Surprisingly, a number of recent ndings suggest that people have difficulty learning in semi- supervised tasks. To further explore this issue, we devised a Category learning task in which the distribution of labeled and unlabeled items suggested alternative organizations of a cate- gory. is design allowed us to determine whether learners combined information from both types of episodes via their patterns of generalization at test. In contrast with the prediction of many models, we nd little evidence that unlabeled items in- uenced categorization behavior when labeled items were also present. Keywords: Semi-supervised Category learning; rule induction; unsupervised learning Can people acquire categories via semi-supervised learning? Despite these arguments, recent empirical attempts to demon- strate semi-supervised Category learning in the lab have met with mixed success. For example, Vandist, De Schryver, and Rosseel (2009) found that adding unlabeled training exam- ples to a mostly supervised task offered no additional bene- t beyond learning from only the supervised trials. However, the Category structures they tested (known as Information- Integration categories) are typically difficult for people to learn even in fully unsupervised settings (Ashby, Queller, & Berretty, 1999), which may explain the limited impact that the unlabeled examples had. On the other hand, Kalish, Rogers, Lang, and Zhu (2011) showed that aer learning a simple Category distinction on a single dimension from a small set of labeled examples, par- ticipants’ estimate of the Category boundary could be shied by the presentation of a large number of unlabeled examples whose distribution was shied compared to the labeled set (see also Lake & McClelland, 2011). While this study provides some evidence of semi-supervised learning, there remain al- ternative explanations of the effect. For example, since the central tendency of both categories are shied in these studies it is unclear whether people are separately updating each cat- egory representation or responding to the global shi in the stimulus space. Finally, Rogers, Kalish, Gibson, Harrison, and Zhu (2010) compared learning in a semi-supervised learning condition with a fully supervised condition. In this study, adding un- labeled items to a supervised Category learning task caused faster learning only when trials were speeded. However, the question of whether people can integrate labeled and unla- beled training examples is logically separate from claims about Introduction Category learning is a critical cognitive ability which is cen- tral to many aspects of cognition. As a result, considerable re- search over the last – years has explored the psychology of Category learning using laboratory tasks. e majority of this work can be divided into two groups. Most research has fo- cused on supervised learning tasks where corrective feedback or Category labels are presented following or alongside each observation of a stimulus (e.g., Medin & Schaffer, 1978; Nosof- sky, 1986). More recently, there has been an interest in unsu- pervised learning, wherein participants must organize exam- ples in the absence of explicit instruction using the distribu- tional properties of the stimuli (e.g., Clapper & Bower, 1994; Love, 2002; Pothos et al., 2011). However, neither of these sit- uations adequately re ect the problem of real world Category learning, in which feedback is not altogether absent nor always present, but is typically sparse and intermittent. Such tasks re- quire learners to combine information from both labeled and unlabeled episodes. In machine learning, this problem is fre- quently studied under the name semi-supervised learning (for review, see Zhu, 2005). Aside from offering a more ecologically relevant approach to the study of Category learning, the study of semi-supervised learning has important implications for theories of human concept learning. Consider the problem of learning a con- crete noun such as horse. One proposal is that word learn- ing essentially links sound tokens (words) to already-acquired hypotheses or representations (Bloom, 2000; Gentner, 1982). Under this view, the label information from a teacher or par- ent about a single example horse must be integrated with the
-
CogSci - Sparse Category labels obstruct generalization of Category Membership
Cognitive Science, 2012Co-Authors: John V Mcdonnell, Carol A Jew, Todd M GureckisAbstract:Sparse Category labels obstruct generalization of Category Membership John V. McDonnell (john.mcdonnell@nyu.edu) Carol A. Jew (carol.jew@nyu.edu) Todd M. Gureckis (todd.gureckis@nyu.edu) New York University, Department of Psychology 6 Washington Place, New York, NY 10003 USA Abstract child’s pre-linguistic grouping of objects in their environment into classes. A similar position is advocated by a number of in uen- tial theories of Category learning which hold that supervised and unsupervised learning are subserved by a single under- lying learning process (e.g., the rational model of categoriza- tion, Anderson, 1991; or the Supervised and Unsupervised STrati ed Adaptive Incremental Network, abbreviated - , Love, Medin, & Gureckis, 2004). Such models naturally predict that semi-supervised learning should not only be pos- sible, but may be the primary way in which people learn cate- gories and their respective names. Studies of human Category learning typically focus on situa- tions where explicit Category labels accompany each example (supervised learning) or on situations were people must infer Category structure entirely from the distribution of unlabeled examples (unsupervised learning). However, real-world cate- gory learning likely involves a mixture of both types of learning (semi-supervised learning). Surprisingly, a number of recent ndings suggest that people have difficulty learning in semi- supervised tasks. To further explore this issue, we devised a Category learning task in which the distribution of labeled and unlabeled items suggested alternative organizations of a cate- gory. is design allowed us to determine whether learners combined information from both types of episodes via their patterns of generalization at test. In contrast with the prediction of many models, we nd little evidence that unlabeled items in- uenced categorization behavior when labeled items were also present. Keywords: Semi-supervised Category learning; rule induction; unsupervised learning Can people acquire categories via semi-supervised learning? Despite these arguments, recent empirical attempts to demon- strate semi-supervised Category learning in the lab have met with mixed success. For example, Vandist, De Schryver, and Rosseel (2009) found that adding unlabeled training exam- ples to a mostly supervised task offered no additional bene- t beyond learning from only the supervised trials. However, the Category structures they tested (known as Information- Integration categories) are typically difficult for people to learn even in fully unsupervised settings (Ashby, Queller, & Berretty, 1999), which may explain the limited impact that the unlabeled examples had. On the other hand, Kalish, Rogers, Lang, and Zhu (2011) showed that aer learning a simple Category distinction on a single dimension from a small set of labeled examples, par- ticipants’ estimate of the Category boundary could be shied by the presentation of a large number of unlabeled examples whose distribution was shied compared to the labeled set (see also Lake & McClelland, 2011). While this study provides some evidence of semi-supervised learning, there remain al- ternative explanations of the effect. For example, since the central tendency of both categories are shied in these studies it is unclear whether people are separately updating each cat- egory representation or responding to the global shi in the stimulus space. Finally, Rogers, Kalish, Gibson, Harrison, and Zhu (2010) compared learning in a semi-supervised learning condition with a fully supervised condition. In this study, adding un- labeled items to a supervised Category learning task caused faster learning only when trials were speeded. However, the question of whether people can integrate labeled and unla- beled training examples is logically separate from claims about Introduction Category learning is a critical cognitive ability which is cen- tral to many aspects of cognition. As a result, considerable re- search over the last – years has explored the psychology of Category learning using laboratory tasks. e majority of this work can be divided into two groups. Most research has fo- cused on supervised learning tasks where corrective feedback or Category labels are presented following or alongside each observation of a stimulus (e.g., Medin & Schaffer, 1978; Nosof- sky, 1986). More recently, there has been an interest in unsu- pervised learning, wherein participants must organize exam- ples in the absence of explicit instruction using the distribu- tional properties of the stimuli (e.g., Clapper & Bower, 1994; Love, 2002; Pothos et al., 2011). However, neither of these sit- uations adequately re ect the problem of real world Category learning, in which feedback is not altogether absent nor always present, but is typically sparse and intermittent. Such tasks re- quire learners to combine information from both labeled and unlabeled episodes. In machine learning, this problem is fre- quently studied under the name semi-supervised learning (for review, see Zhu, 2005). Aside from offering a more ecologically relevant approach to the study of Category learning, the study of semi-supervised learning has important implications for theories of human concept learning. Consider the problem of learning a con- crete noun such as horse. One proposal is that word learn- ing essentially links sound tokens (words) to already-acquired hypotheses or representations (Bloom, 2000; Gentner, 1982). Under this view, the label information from a teacher or par- ent about a single example horse must be integrated with the
Bradley C Love - One of the best experts on this subject based on the ideXlab platform.
-
abstract neural representations of Category Membership beyond information coding stimulus or response
Journal of Cognitive Neuroscience, 2020Co-Authors: Robert M Mok, Bradley C LoveAbstract:For decades, researchers have debated whether mental representations are symbolic or grounded in sensory inputs and motor programs. Certainly, aspects of mental representations are grounded. Howeve...
-
abstract neural representations of Category Membership beyond information coding stimulus or response
bioRxiv, 2020Co-Authors: Robert M Mok, Bradley C LoveAbstract:For decades, researchers have debated whether mental representations are symbolic or grounded in sensory inputs and motor programs. Certainly, aspects of mental representations are grounded. However, does the brain also contain abstract concept representations that mediate between perception and action in a flexible manner not tied to the details of sensory inputs and motor programs? Such conceptual pointers would be useful when concept remain constant despite changes in appearance and associated actions. We evaluated whether human participants acquire such representations using functional magnetic resonance imaging (fMRI). Participants completed a probabilistic concept learning task in which sensory, motor, and Category variables were not perfectly coupled nor entirely independent, making it possible to observe evidence for abstract representations or purely grounded representations. To assess how the learned concept structure is represented in the brain, we examined brain regions implicated in flexible cognition (e.g., prefrontal and parietal cortex) that are most likely to encode an abstract representation removed from sensory-motor details. We also examined sensory-motor regions that might encode grounded sensory-motor based representations tuned for categorization. Using a cognitive model to estimate participants9 Category rule and multivariate pattern analysis of fMRI data, we found left prefrontal cortex and MT coded for Category in absence of information coding for stimulus or response. Because Category was based on the stimulus, finding an abstract representation of Category was not inevitable. Our results suggest that certain brain areas support categorization behaviour by constructing concept representations in a format akin to a symbol that differs from stimulus-motor codes.
Daniel M. Weiskopf - One of the best experts on this subject based on the ideXlab platform.
-
An electrophysiological comparison of visual categorization and recognition memory
Cognitive Affective & Behavioral Neuroscience, 2002Co-Authors: Tim Curran, James W. Tanaka, Daniel M. WeiskopfAbstract:Object categorization emphasizes the similarities that bind exemplars into categories, whereas recognition memory emphasizes the specific identification of previously encountered exemplars. Mathematical modeling has highlighted similarities in the computational requirements of these tasks, but neuropsychological research has suggested that categorization and recognition may depend on separate brain systems. Following training with families of novel visual shapes ( blobs ), event-related brain potentials (ERPs) were recorded during both categorization and recognition tasks. ERPs related to early visual processing (N1, 156–200 msec) were sensitive to Category Membership. Middle latency ERPs (FN400 effects, 300–500 msec) were sensitive to both Category Membership and old/new differences. Later ERPs (parietal effects, 400–800 msec) were primarily affected by old/new differences. Thus, there was a temporal transition so that earlier processes were more sensitive to categorical discrimination and later processes were more sensitive to recognition-related discrimination. Aspects of these results are consistent with both mathematical modeling and neuropsychological perspectives.
-
An electrophysiological comparison of visual categorization and recognition memory
Cognitive Affective & Behavioral Neuroscience, 2002Co-Authors: Tim Curran, James W. Tanaka, Daniel M. WeiskopfAbstract:Object categorization emphasizes the similarities that bind exemplars into categories, whereas recognition memory emphasizes the specific identification of previously encountered exemplars. Mathematical modeling has highlighted similarities in the computational requirements of these tasks, but neuropsychological research has suggested that categorization and recognition may depend on separate brain systems. Following training with families of novel visual shapes ( blobs ), event-related brain potentials (ERPs) were recorded during both categorization and recognition tasks. ERPs related to early visual processing (N1, 156–200 msec) were sensitive to Category Membership. Middle latency ERPs (FN400 effects, 300–500 msec) were sensitive to both Category Membership and old/new differences. Later ERPs (parietal effects, 400–800 msec) were primarily affected by old/new differences. Thus, there was a temporal transition so that earlier processes were more sensitive to categorical discrimination and later processes were more sensitive to recognition-related discrimination. Aspects of these results are consistent with both mathematical modeling and neuropsychological perspectives.
