File Download

There are no files associated with this item.

  Links for fulltext
     (May Require Subscription)
Supplementary

Conference Paper: Eye Movement Patterns in Face Recognition are Associated with Cognitive Decline in Older Adults: An HMM Approach.

TitleEye Movement Patterns in Face Recognition are Associated with Cognitive Decline in Older Adults: An HMM Approach.
Authors
Issue Date2018
PublisherAssociation for Research in Vision and Ophthalmology. The Journal's web site is located at http://wwwjournalofvisionorg/
Citation
Vision Sciences Society (VSS) 18th Annual Meeting, St. Pete Beach, Florida, USA, 18-23 May 2018. Meeting Abstracts in Journal of Vision, 2018, v. 18 n. 10, abstract no. 231 How to Cite?
AbstractCurrent methods for early identification of neurodegenerative cognitive decline are typically based on neuroimaging technologies, which are expertise-demanding and not commonly available. Here we examined the potential use of eye tracking as an easily deployable and inexpensive screening tool through the Eye Movement analysis with Hidden Markov Models (EMHMM) approach. This approach summarizes a participant's eye movements in a visual task with person-specific regions of interest (ROIs) and transition probabilities among the ROIs in an HMM. Individual HMMs can be clustered to discover common patterns. Similarity between an individual pattern and a discovered common pattern can be quantitatively assessed using log-likelihood measures. This measure can be used to examine the relationship between eye movement patterns and cognitive performance. In experiment 1, we recruited young and older adults to perform a face recognition task and discovered 'holistic' (mostly fixating around the face center) and 'analytic' (frequent transitions among the two eyes and the face center) patterns. Significantly more older participants adopted the holistic pattern, whilst more young participants adopted the analytic pattern. The analytic pattern yielded better face recognition performance regardless of age. Importantly, older participants' cognitive status, as assessed by the Montreal Cognitive Assessment, was negatively correlated with eye movement similarity to the holistic pattern. Experiment 2 examined whether the holistic and analytic patterns discovered in Experiment 1 could be used to assess eye movements of new participants for screening purposes. New older participants performed the same face recognition task with different stimuli. Consistent with Experiment 1, high similarity to the holistic pattern was correlated with low cognitive status, particularly in executive and visual attention functioning. This result demonstrates well the power of the EMHMM approach, suggesting the possibility of using eye movements in a simple visual task as a fast, easily deployable, and inexpensive screening tool for cognitive decline.
Persistent Identifierhttp://hdl.handle.net/10722/265196
ISSN
2023 Impact Factor: 2.0
2023 SCImago Journal Rankings: 0.849

 

DC FieldValueLanguage
dc.contributor.authorChan, YH-
dc.contributor.authorChan, AB-
dc.contributor.authorLee, TMC-
dc.contributor.authorHsiao, JHW-
dc.date.accessioned2018-11-20T02:01:59Z-
dc.date.available2018-11-20T02:01:59Z-
dc.date.issued2018-
dc.identifier.citationVision Sciences Society (VSS) 18th Annual Meeting, St. Pete Beach, Florida, USA, 18-23 May 2018. Meeting Abstracts in Journal of Vision, 2018, v. 18 n. 10, abstract no. 231-
dc.identifier.issn1534-7362-
dc.identifier.urihttp://hdl.handle.net/10722/265196-
dc.description.abstractCurrent methods for early identification of neurodegenerative cognitive decline are typically based on neuroimaging technologies, which are expertise-demanding and not commonly available. Here we examined the potential use of eye tracking as an easily deployable and inexpensive screening tool through the Eye Movement analysis with Hidden Markov Models (EMHMM) approach. This approach summarizes a participant's eye movements in a visual task with person-specific regions of interest (ROIs) and transition probabilities among the ROIs in an HMM. Individual HMMs can be clustered to discover common patterns. Similarity between an individual pattern and a discovered common pattern can be quantitatively assessed using log-likelihood measures. This measure can be used to examine the relationship between eye movement patterns and cognitive performance. In experiment 1, we recruited young and older adults to perform a face recognition task and discovered 'holistic' (mostly fixating around the face center) and 'analytic' (frequent transitions among the two eyes and the face center) patterns. Significantly more older participants adopted the holistic pattern, whilst more young participants adopted the analytic pattern. The analytic pattern yielded better face recognition performance regardless of age. Importantly, older participants' cognitive status, as assessed by the Montreal Cognitive Assessment, was negatively correlated with eye movement similarity to the holistic pattern. Experiment 2 examined whether the holistic and analytic patterns discovered in Experiment 1 could be used to assess eye movements of new participants for screening purposes. New older participants performed the same face recognition task with different stimuli. Consistent with Experiment 1, high similarity to the holistic pattern was correlated with low cognitive status, particularly in executive and visual attention functioning. This result demonstrates well the power of the EMHMM approach, suggesting the possibility of using eye movements in a simple visual task as a fast, easily deployable, and inexpensive screening tool for cognitive decline.-
dc.languageeng-
dc.publisherAssociation for Research in Vision and Ophthalmology. The Journal's web site is located at http://wwwjournalofvisionorg/-
dc.relation.ispartofJournal of Vision-
dc.relation.ispartofAnnual Meeting of Vision Sciences Society (VSS)-
dc.titleEye Movement Patterns in Face Recognition are Associated with Cognitive Decline in Older Adults: An HMM Approach.-
dc.typeConference_Paper-
dc.identifier.emailLee, TMC: tmclee@hku.hk-
dc.identifier.emailHsiao, JHW: jhsiao@hku.hk-
dc.identifier.authorityLee, TMC=rp00564-
dc.identifier.authorityHsiao, JHW=rp00632-
dc.description.naturelink_to_OA_fulltext-
dc.identifier.doi10.1167/18.10.231-
dc.identifier.hkuros295947-
dc.identifier.volume18-
dc.identifier.issue10-
dc.identifier.spage231-
dc.identifier.epage231-
dc.publisher.placeUnited States-
dc.identifier.issnl1534-7362-

Export via OAI-PMH Interface in XML Formats


OR


Export to Other Non-XML Formats