Optimizing Feedback-based Learning in Children With Developmental Language Disorder

Summary

This project aims to optimize a critical but understudied ingredient of language intervention provided to children with developmental language disorder (DLD) - feedback. The project will bridge a gap between previous findings in our lab of inefficient feedback processing in DLD and clinical practice by identifying the conditions under which feedback-based learning can be improved in DLD. The investigators hypothesize that the effectiveness of feedback can be significantly enhanced for children with DLD when it is tailored to their unique learning strengths. The rationale for this project is based on evidence that feedback-based learning can be improved by enhancing the dominance of an intact learning system. The project will achieve its aim by manipulating (1) the timing of the feedback (immediate vs. delayed) and (2) the level of the learner's involvement in error correction dictated by feedback (active vs. passive correction). Aim 1 will determine the effect of manipulating feedback timing on learning in 140 school-age children (8-12 years) with DLD. While immediate feedback is processed by the striatum, which is also implicated in implicit learning, delaying the feedback by a few seconds shifts feedback processing to the mediate temporal lobe (MTL)-based declarative learning system. Evidence that delaying feedback improves learning in DLD would support the hypothesis of the implicit deficit theory that intervention should capitalize on declarative learning mechanisms. The project will test a novel alternative feedback-learning parity hypothesis whereby feedback-based learning is optimized when the timing of the feedback is aligned with the dominant learning system at a given time (i.e., immediate feedback during striatal-based probabilistic learning; delayed feedback during MTL-based declarative learning). Within the same group of children, Aim 2 will compare feedback-based learning in children with DLD when feedback (a) prompts active self-correction or (b) passively exposes learners to error corrections (corrective recast). Children will engage in two nonword-object paired-associate learning tasks. In one task, feedback will promote active self-correction, which is in line with declarative learning. In the other task, feedback will passively expose the learner to corrective feedback in a manner consistent with teaching approaches aiming at reducing awareness of errors. The project will determine whether children with DLD learn better when feedback prompts self-correction or when they are exposed to passive corrections. Electrophysiological measures will indicate whether passive corrections (corrective recast) are processed as negative feedback by children with DLD. For both aims, behavioral indicators of response to feedback will be complemented by electrophysiological measures of feedback processing that can determine the involvement of the striatum and MTL brain systems during the learning process. This work is scientifically and clinically significant because elucidating what manipulations optimize feedback-based learning will enhance our understanding of the impaired learning mechanism in DLD and will provide clinical guidance on what type of feedback to use during an intervention.

Conditions

• Developmental Language Disorder

Interventions

BEHAVIORAL

Declarative learning with immediate feedback

Children will complete a paired associate learning task on a computer. Each response on half of the blocks will be followed by performance feedback. Feedback will be provided 500 ms following a response (button press)

BEHAVIORAL

Declarative learning with delayed feedback

Children will complete a paired associate learning task on a computer. Each response on half of the blocks will be followed by performance feedback. Feedback will be provided 5000 ms following a response (button press)

BEHAVIORAL

Probabilistic learning with immediate feedback

Children will complete a probabilistic learning task on a computer. Each response on half of the blocks will be followed by feedback. Feedback will be provided 500 ms following a response (button press)

BEHAVIORAL

Probabilistic learning with delayed feedback

Children will complete a probabilistic learning task on a computer. Each response on half of the blocks will be followed by feedback. Feedback will be provided 5000 ms following a response (button press)

BEHAVIORAL

Word learning task with active feedback

Children will complete a two-choice nonword learning task in which they will be tasked with learning to associate novel objects with novel nonword names through trial and error guided by feedback. Each response will be followed by feedback either indicating that the response is correct or prompting the child to try again (i.e., to correct an error)

BEHAVIORAL

Word learning task with passive feedback

Children will complete a two-choice nonword learning task in which they will be tasked with learning to associate novel objects with novel nonword names. Regardless of response accuracy on each trial, each response will be followed by a visual presentation of the correct response.

Sponsors & Collaborators

• MGH Institute of Health Professions

  lead OTHER

Principal Investigators

• Yael Arbel, PhD · MGH Institute of Health Professions

Study Design

Allocation

NA

Purpose

BASIC_SCIENCE

Masking

NONE

Model

SINGLE_GROUP

Eligibility

Min Age

8 Years

Max Age

12 Years

Sex

ALL

Healthy Volunteers

No

Timeline & Regulatory

Start

2023-07-20

Primary Completion

2027-10-01

Completion

2027-10-01

Countries

• United States

Study Locations