Repository logo
Log In(current)
  1. Home
  2. Colleges & Schools
  3. Graduate School
  4. Doctoral Dissertations
  5. DECODING INFANT CATEGORIZATION OF OTHER-SPECIES FACES USING MACHINE LEARNING METHODS
Details

DECODING INFANT CATEGORIZATION OF OTHER-SPECIES FACES USING MACHINE LEARNING METHODS

Date Issued
August 2026
Author(s)
Clayton, Kenna Reed Harmon
The University of Tennessee
Advisor(s)
Greg Reynolds
Additional Advisor(s)
Aaron Buss
Haileab Hilafu
David Stutterer
Permanent URI
https://trace.tennessee.edu/handle/20.500.14382/61874
Abstract
Infants initially exhibit broad perceptual sensitivity to visual stimuli that becomes increasingly specialized with age and experience in a developmental process described as perceptual narrowing (Maurer & Werker, 2014). In face processing, this transition is reflected in a decline in infants’ ability to individuate non-human primate faces across the first year of life in the absence of continued exposure (Pascalis et al., 2002; Pascalis et al., 2005). One account proposes that perceptual narrowing reflects experience-dependent changes in learned attention (Simpson et al., 2014), such that frequently encountered faces are processed at the individual level whereas less familiar faces are processed more categorically (Hugenberg et al., 2010; Nelson, 2001; Reynolds & Roth, 2018). Experimental evidence further suggests that exposure to multiple exemplars facilitates subordinate-level categorization of other-species faces relative to exposure to a single exemplar (Dixon et al., 2019). The present study examined how age and initial learning conditions influence infants’neural representations of other-species faces (Capuchin and Macaque monkeys). 6- and 12-month-old infants were familiarized with either a single monkey face exemplar or multiple exemplars from one species before viewing familiar faces, novel faces from the same species, and novel faces from a different species during EEG recording. In addition to traditional event-related potential (ERP) analyses of components associated with infant face processing and attention (de Haan, 2007; Halit et al., 2003), time-resolved multivariate pattern classification (MVPA) and representational geometry analyses were used to assess whether distributed neural activity patterns supported subordinate-level category discrimination across development and learning condition (Bayet et al., 2020; Grootswagers et al., 2017). ERP analyses revealed age-related modulation of face-sensitive components, with 6-month-old infants showing evidence of identity-level differentiation across stimulus types and reduced neural differentiation among monkey faces at 12 months consistent with perceptual narrowing. In contrast, multivariate analyses revealed reliable subordinate-level classification of monkey species at 6 months, with temporally stable category representations emerging primarily following single-exemplar familiarization. These findings indicate that perceptual narrowing reflects experience-dependent reorganization of representational space across infancy and provide support for the importance of multivariate neural approaches for revealing early category structure not detectable using traditional ERP analyses alone.
Subjects

perceptual narrowing

Psychology, Developme...

Degree
Doctor of Philosophy
Major
Experimental Psychology
File(s)
Thumbnail Image
Name

CLAYTON-PRIMARY-2026.pdf

Size

4.78 MB

Format

Adobe PDF

Checksum (MD5)

cc1fb3518e9168c9854328dfc1945481


University Libraries

1015 Volunteer Boulevard
Knoxville, TN 37996
865-974-4351

Map & Directions
Donate to the Libraries
  • About
  • John C. Hodges Society
  • Speaking Volumes magazine
  • Outreach
  • Directory
  • Employment
  • Policies
  • Library Intranet
University of Tennessee power T logo

The University of Tennessee, Knoxville
Knoxville, Tennessee 37996
865-974-1000

Events
A-Z
Apply
Privacy
Map
Directory
Give to UT
Accessibility

Built with DSpace-CRIS software - Extension maintained and optimized by 4Science