Sensory cue-combination in the context of newly learned categories

Kaitlyn R Bankieris,Richard N Aslin,Vikranth Rao Bejjanki

doi:10.1038/s41598-017-11341-7

Kaitlyn R Bankieris, Richard N Aslin + Show 1 more

Open Access

https://doi.org/10.1038/s41598-017-11341-7

Copy DOI

Journal: Scientific Reports	Publication Date: Sep 7, 2017
Citations: 8	License type: open-access

Affiliation: University of Rochester, Hamilton College

Abstract

A large body of prior research has evaluated how humans combine multiple sources of information pertaining to stimuli drawn from continuous dimensions, such as distance or size. These prior studies have repeatedly demonstrated that in these circumstances humans integrate cues in a near-optimal fashion, weighting cues according to their reliability. However, most of our interactions with sensory information are in the context of categories such as objects and phonemes, thereby requiring a solution to the cue combination problem by mapping sensory estimates from continuous dimensions onto task-relevant categories. Previous studies have examined cue combination with natural categories (e.g., phonemes), providing qualitative evidence that human observers utilize information about the distributional properties of task-relevant categories, in addition to sensory information, in such categorical cue combination tasks. In the present study, we created and taught human participants novel audiovisual categories, thus allowing us to quantitatively evaluate participants’ integration of sensory and categorical information. Comparing participant behavior to the predictions of a statistically optimal observer that ideally combines all available sources of information, we provide the first evidence, to our knowledge, that human observers combine sensory and category information in a statistically optimal manner.

Highlights

A large body of prior research has evaluated how humans combine multiple sources of information pertaining to stimuli drawn from continuous dimensions, such as distance or size
A large body of prior research has evaluated the mechanism of sensory cue-combination across continuous dimensions such as size
Studies evaluating cue combination across continuous dimensions have demonstrated that humans do integrate multiple sources of information efficiently, following this statistically optimal strategy of weighting sensory cues based on their reliability

Summary

Introduction

A large body of prior research has evaluated how humans combine multiple sources of information pertaining to stimuli drawn from continuous dimensions, such as distance or size These prior studies have repeatedly demonstrated that in these circumstances humans integrate cues in a near-optimal fashion, weighting cues according to their reliability. The combined estimate of the stimulus is a weighted linear combination of the estimates suggested by the two sensory signals: www.nature.com/scientificreports/ These equations (1–3) describe the behaviour of an ideal observer when combining two cues lying along continuous dimensions for a given sensory stimulus, such as spatial location or size, because this approach minimizes the variance of the resulting estimate[5]. Studies evaluating cue combination across continuous dimensions have demonstrated that humans do integrate multiple sources of information efficiently, following this statistically optimal strategy of weighting sensory cues based on their reliability (e.g., refs 2, 5–12). We seek to bridge the gap between the cue integration literature – which examines how cues with sensory variance are combined – and an extensive body of categorisation literature which investigates how features (with negligible sensory noise) are combined under category uncertainty (e.g., refs 16 and 17)

Methods

Results

Conclusion