Unimodal statistical learning produces multimodal object-like representations
The concept of objects is fundamental to cognition and is defined by a consistent set of sensory properties and physical affordances. Although it is unknown how the abstract concept of an object emerges, most accounts assume that visual or haptic boundaries are crucial in this process. Here, we test...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
eLife Sciences Publications Ltd
2019-05-01
|
| Series: | eLife |
| Subjects: | |
| Online Access: | https://elifesciences.org/articles/43942 |
| _version_ | 1811199796011073536 |
|---|---|
| author | Gábor Lengyel Goda Žalalytė Alexandros Pantelides James N Ingram József Fiser Máté Lengyel Daniel M Wolpert |
| author_facet | Gábor Lengyel Goda Žalalytė Alexandros Pantelides James N Ingram József Fiser Máté Lengyel Daniel M Wolpert |
| author_sort | Gábor Lengyel |
| collection | DOAJ |
| description | The concept of objects is fundamental to cognition and is defined by a consistent set of sensory properties and physical affordances. Although it is unknown how the abstract concept of an object emerges, most accounts assume that visual or haptic boundaries are crucial in this process. Here, we tested an alternative hypothesis that boundaries are not essential but simply reflect a more fundamental principle: consistent visual or haptic statistical properties. Using a novel visuo-haptic statistical learning paradigm, we familiarised participants with objects defined solely by across-scene statistics provided either visually or through physical interactions. We then tested them on both a visual familiarity and a haptic pulling task, thus measuring both within-modality learning and across-modality generalisation. Participants showed strong within-modality learning and ‘zero-shot’ across-modality generalisation which were highly correlated. Our results demonstrate that humans can segment scenes into objects, without any explicit boundary cues, using purely statistical information. |
| first_indexed | 2024-04-12T01:53:28Z |
| format | Article |
| id | doaj.art-26e7d9ec734742fe91ee8a7406a49219 |
| institution | Directory Open Access Journal |
| issn | 2050-084X |
| language | English |
| last_indexed | 2024-04-12T01:53:28Z |
| publishDate | 2019-05-01 |
| publisher | eLife Sciences Publications Ltd |
| record_format | Article |
| series | eLife |
| spelling | doaj.art-26e7d9ec734742fe91ee8a7406a492192022-12-22T03:52:51ZengeLife Sciences Publications LtdeLife2050-084X2019-05-01810.7554/eLife.43942Unimodal statistical learning produces multimodal object-like representationsGábor Lengyel0https://orcid.org/0000-0002-1535-3250Goda Žalalytė1https://orcid.org/0000-0002-0012-9950Alexandros Pantelides2https://orcid.org/0000-0002-6234-6061James N Ingram3https://orcid.org/0000-0003-2567-504XJózsef Fiser4https://orcid.org/0000-0002-7064-0690Máté Lengyel5https://orcid.org/0000-0001-7266-0049Daniel M Wolpert6https://orcid.org/0000-0003-2011-2790Department of Cognitive Science, Central European University, Budapest, HungaryComputational and Biological Learning Lab, Department of Engineering, University of Cambridge, Cambridge, United KingdomComputational and Biological Learning Lab, Department of Engineering, University of Cambridge, Cambridge, United KingdomComputational and Biological Learning Lab, Department of Engineering, University of Cambridge, Cambridge, United Kingdom; Zuckerman Mind Brain Behavior Institute, Department of Neuroscience, Columbia University, New York, United StatesDepartment of Cognitive Science, Central European University, Budapest, HungaryDepartment of Cognitive Science, Central European University, Budapest, Hungary; Computational and Biological Learning Lab, Department of Engineering, University of Cambridge, Cambridge, United KingdomComputational and Biological Learning Lab, Department of Engineering, University of Cambridge, Cambridge, United Kingdom; Zuckerman Mind Brain Behavior Institute, Department of Neuroscience, Columbia University, New York, United StatesThe concept of objects is fundamental to cognition and is defined by a consistent set of sensory properties and physical affordances. Although it is unknown how the abstract concept of an object emerges, most accounts assume that visual or haptic boundaries are crucial in this process. Here, we tested an alternative hypothesis that boundaries are not essential but simply reflect a more fundamental principle: consistent visual or haptic statistical properties. Using a novel visuo-haptic statistical learning paradigm, we familiarised participants with objects defined solely by across-scene statistics provided either visually or through physical interactions. We then tested them on both a visual familiarity and a haptic pulling task, thus measuring both within-modality learning and across-modality generalisation. Participants showed strong within-modality learning and ‘zero-shot’ across-modality generalisation which were highly correlated. Our results demonstrate that humans can segment scenes into objects, without any explicit boundary cues, using purely statistical information.https://elifesciences.org/articles/43942statistical learningvisual statistical learninghaptic statistical learningobject representationszero-shot generalization |
| spellingShingle | Gábor Lengyel Goda Žalalytė Alexandros Pantelides James N Ingram József Fiser Máté Lengyel Daniel M Wolpert Unimodal statistical learning produces multimodal object-like representations eLife statistical learning visual statistical learning haptic statistical learning object representations zero-shot generalization |
| title | Unimodal statistical learning produces multimodal object-like representations |
| title_full | Unimodal statistical learning produces multimodal object-like representations |
| title_fullStr | Unimodal statistical learning produces multimodal object-like representations |
| title_full_unstemmed | Unimodal statistical learning produces multimodal object-like representations |
| title_short | Unimodal statistical learning produces multimodal object-like representations |
| title_sort | unimodal statistical learning produces multimodal object like representations |
| topic | statistical learning visual statistical learning haptic statistical learning object representations zero-shot generalization |
| url | https://elifesciences.org/articles/43942 |
| work_keys_str_mv | AT gaborlengyel unimodalstatisticallearningproducesmultimodalobjectlikerepresentations AT godazalalyte unimodalstatisticallearningproducesmultimodalobjectlikerepresentations AT alexandrospantelides unimodalstatisticallearningproducesmultimodalobjectlikerepresentations AT jamesningram unimodalstatisticallearningproducesmultimodalobjectlikerepresentations AT jozseffiser unimodalstatisticallearningproducesmultimodalobjectlikerepresentations AT matelengyel unimodalstatisticallearningproducesmultimodalobjectlikerepresentations AT danielmwolpert unimodalstatisticallearningproducesmultimodalobjectlikerepresentations |