Origins of Infant Word Learning: Object Recognition
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
In the first months of life, infants are better able to remember and categorize objects that occur with words, suggesting that language and vision are linked in early human development. However, it is largely unknown how and why learning a language influences how infants see the world. This project aims to discover how words influence how 6- to 18-month-old infants see objects and how this process ultimately leads to word learning. In order to understand this process, we must first understand the recognition of objects in early infancy. Masking studies have offered insight into the influence of feedback processing and recurrence during visual recognition. Within these studies, a stimulus is believed to be rendered "invisible" as a direct result of backward masking--where a mask is presented or remains after a target stimulus (Nakashima et al., 2021; Kotsoni et al., 2006). One theory indicates this is due to the disruption of the feedback processing of a stimulus, in order for a stimulus to be recognised (Silverstein, 2015). This is not to say the visual system does not notice a visual input, but that after the visual system has detected this visual input, the ability to discriminate between the stimuli and the background and detect what the stimuli is becomes disrupted. Further, object substitution masking studies (a type of backward masking where the mask appears around the target and remains on screen after the target disappears) have indicated this feedback information may be assisted by recurrent pathways allowing information to loop in order to be processed (Jannati, Spalek, Di Lollo, 2013). Here, the mask is thought to result in a mismatch between the neural activity triggered by the initial display which ascends to higher brain regions, and the ongoing low-level activity attempting to process the stimulus. The neural activity in the higher brain regions proposes a perceptual hypothesis that loops down to the lower level to correlate with the ongoing lower-level activity (Di Lollo, Enns, & Rensink, 2000). Thus, object substitution masking may result in a mismatch due to uneven levels of activity between this perceptual hypothesis (the reentrant hypothesis) and the ongoing lower-level activity (Jannati, Spalek, & Di Lollo, 2013). However, another view proposes backward masking interferes with the target through feedforward lateral inhibition, preventing further processing (Macknick, 2006). Silverstein (2015) suggests, however, that masking may occur as a result of both lateral inhibition and disruption to recurrent processing. Biophysical modelling indicated a mask that provided enough lateral inhibition to disrupt the processing of a stimuli, which was also prevented from being reinforced by recurrent loops, was unable to pass through consecutive visual areas in order to be successfully detected and recognised. The majority of research examining the influence of this masking on visual recognition has only been conducted on mature pathways with adults or adult-like models, with little known about how masking could influence the developing visual processing system in infancy. Nakashima et al., (2021) demonstrated the presence of recurrence on 8 month old infants during facial recognition on an object substitution masking task. Looking to a face was significantly lower when a mask consisting of four dots around the face remained on the screen (delayed offset) after the face disappeared than when the dots disappeared simultaneously with the face (simultaneous offset). However, there was no difference in looking between a delayed and a simultaneous offset of the mask at 3 to 6 months of age. It was concluded that by 8 months of age, recurrent pathways may have matured to a level where recurrent loops can become disrupted by the mask. This provides support for the feedback processing theory, as these findings are unlikely the result of 3 to 6 month old infants having faster feedforward processing, resulting in the mask having no effect, than older infants and even adults. It is likely that feedforward processing is used to interpret visual stimuli in the first 6 months of life, but as recurrent pathways mature and visual information becomes familiar, feedback information is used. The early development of these processes is still relatively unknown during infancy, and this process has not been examined using eye tracking methods which may allow us to examine infant responses in more detail. Importantly, these conclusions must be examined in the context of word learning. Objects are rarely presented to infants in isolation. Labels can aid visual processing, for example Lupyan and Ward (2013) found that the use of labels assisted the detection of objects which were suppressed from perception via continuous flash suppression. Specifically, the use of valid labels lead to object recognition, whereas the use of invalid labels did not. Consequently, it is important we understand the use of feedback processing and recurrence in object recognition during infancy, when being familiarised with objects is often paired with labels. To examine this, we should first explore how object substitution masking influences object recognition in infancy. Given Nakashima et al., (2021) found object substitution masking to be successful from 7 to 8 month olds, we will test two age groups beginning at 7 to 9 months. The second age group will be 13 to 15 months, as these infants should have more familiarity with the objects and their labels. Understanding the effects of object substitution masking at these ages will then allow us to go on to explore the influence of adding words in the form of object labels, and how words may reduce the effects of object substitution masking.
Fields of Study & Subjects
Language Development and Disorders
Child and Animal Learning Development
Categorization
perception
and language
Publication details
- Journal / source
- OSF Preprints (OSF Preprints)
- Publication date
- 25 March 2026
- Language
- EN
- Document type
- Preprint