Limits...
Fragment-based learning of visual object categories in non-human primates.

Kromrey S, Maestri M, Hauffen K, Bart E, Hegdé J - PLoS ONE (2010)

Bottom Line: Recent research has shown that the visual system can use local, informative image fragments of a given object, rather than the whole object, to classify it into a familiar category.We have previously reported, using human psychophysical studies, that when subjects learn new object categories using whole objects, they incidentally learn informative fragments, even when not required to do so.However, the neuronal mechanisms by which we acquire and use informative fragments, as well as category knowledge itself, have remained unclear.

View Article: PubMed Central - PubMed

Affiliation: Brain and Behavior Discovery Institute, Medical College of Georgia, Augusta, Georgia, USA.

ABSTRACT
When we perceive a visual object, we implicitly or explicitly associate it with an object category we know. Recent research has shown that the visual system can use local, informative image fragments of a given object, rather than the whole object, to classify it into a familiar category. We have previously reported, using human psychophysical studies, that when subjects learn new object categories using whole objects, they incidentally learn informative fragments, even when not required to do so. However, the neuronal mechanisms by which we acquire and use informative fragments, as well as category knowledge itself, have remained unclear. Here we describe the methods by which we adapted the relevant human psychophysical methods to awake, behaving monkeys and replicated key previous psychophysical results. This establishes awake, behaving monkeys as a useful system for future neurophysiological studies not only of informative fragments in particular, but also of object categorization and category learning in general.

Show MeSH
Performance over the course of testing in Experiment 2.The performance of the animals during the testing sessions is shown. Each data point is averaged from 120 trials from each animal (60 trials each of Main- and Control fragments), randomly interleaved with each other. Note that the performance is lower than the average response for Main fragments (Fig. 7A), because the responses were averaged across Main- and Control fragments. Similar results were obtained when the data were analyzed separately for Main-, Control- or IPControl fragments in this experiment and in Experiment 1 (3-way ANOVA, testing blocks x fragment number x fragment type; p<0.05 for testing blocks).
© Copyright Policy
Related In: Results  -  Collection


getmorefigures.php?uid=PMC2991334&req=5

pone-0015444-g008: Performance over the course of testing in Experiment 2.The performance of the animals during the testing sessions is shown. Each data point is averaged from 120 trials from each animal (60 trials each of Main- and Control fragments), randomly interleaved with each other. Note that the performance is lower than the average response for Main fragments (Fig. 7A), because the responses were averaged across Main- and Control fragments. Similar results were obtained when the data were analyzed separately for Main-, Control- or IPControl fragments in this experiment and in Experiment 1 (3-way ANOVA, testing blocks x fragment number x fragment type; p<0.05 for testing blocks).

Mentions: One potential concern about the above results is that the animals may have learned the fragments during the testing itself, because they were rewarded for correct responses. If this is the case, one would expect that (a) the performance would be at or near chance levels at the beginning of testing, since the animals had not encountered the fragments before, and (b) the performance would improve during the course of testing, as the animals presumably learned the fragments based on feedback implicit in the rewards. An examination of the testing data showed that neither of the above scenarios was applicable to our data (Fig. 8). The performance was above chance levels and stable throughout the testing (binomial proportions test, p>0.05). These results are consistent with a scenario where the animals learned the fragments to asymptotic levels during training with whole objects, and therefore rewarding them during testing did not result in any further learning.


Fragment-based learning of visual object categories in non-human primates.

Kromrey S, Maestri M, Hauffen K, Bart E, Hegdé J - PLoS ONE (2010)

Performance over the course of testing in Experiment 2.The performance of the animals during the testing sessions is shown. Each data point is averaged from 120 trials from each animal (60 trials each of Main- and Control fragments), randomly interleaved with each other. Note that the performance is lower than the average response for Main fragments (Fig. 7A), because the responses were averaged across Main- and Control fragments. Similar results were obtained when the data were analyzed separately for Main-, Control- or IPControl fragments in this experiment and in Experiment 1 (3-way ANOVA, testing blocks x fragment number x fragment type; p<0.05 for testing blocks).
© Copyright Policy
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC2991334&req=5

pone-0015444-g008: Performance over the course of testing in Experiment 2.The performance of the animals during the testing sessions is shown. Each data point is averaged from 120 trials from each animal (60 trials each of Main- and Control fragments), randomly interleaved with each other. Note that the performance is lower than the average response for Main fragments (Fig. 7A), because the responses were averaged across Main- and Control fragments. Similar results were obtained when the data were analyzed separately for Main-, Control- or IPControl fragments in this experiment and in Experiment 1 (3-way ANOVA, testing blocks x fragment number x fragment type; p<0.05 for testing blocks).
Mentions: One potential concern about the above results is that the animals may have learned the fragments during the testing itself, because they were rewarded for correct responses. If this is the case, one would expect that (a) the performance would be at or near chance levels at the beginning of testing, since the animals had not encountered the fragments before, and (b) the performance would improve during the course of testing, as the animals presumably learned the fragments based on feedback implicit in the rewards. An examination of the testing data showed that neither of the above scenarios was applicable to our data (Fig. 8). The performance was above chance levels and stable throughout the testing (binomial proportions test, p>0.05). These results are consistent with a scenario where the animals learned the fragments to asymptotic levels during training with whole objects, and therefore rewarding them during testing did not result in any further learning.

Bottom Line: Recent research has shown that the visual system can use local, informative image fragments of a given object, rather than the whole object, to classify it into a familiar category.We have previously reported, using human psychophysical studies, that when subjects learn new object categories using whole objects, they incidentally learn informative fragments, even when not required to do so.However, the neuronal mechanisms by which we acquire and use informative fragments, as well as category knowledge itself, have remained unclear.

View Article: PubMed Central - PubMed

Affiliation: Brain and Behavior Discovery Institute, Medical College of Georgia, Augusta, Georgia, USA.

ABSTRACT
When we perceive a visual object, we implicitly or explicitly associate it with an object category we know. Recent research has shown that the visual system can use local, informative image fragments of a given object, rather than the whole object, to classify it into a familiar category. We have previously reported, using human psychophysical studies, that when subjects learn new object categories using whole objects, they incidentally learn informative fragments, even when not required to do so. However, the neuronal mechanisms by which we acquire and use informative fragments, as well as category knowledge itself, have remained unclear. Here we describe the methods by which we adapted the relevant human psychophysical methods to awake, behaving monkeys and replicated key previous psychophysical results. This establishes awake, behaving monkeys as a useful system for future neurophysiological studies not only of informative fragments in particular, but also of object categorization and category learning in general.

Show MeSH