Systematic Review of Matrix Training for Individuals With Autism Spectrum Disorder

Abstract

Matrix training is an intervention that systematically programs for the occurrence of generative outcomes, mostly in the area of language. A literature search identified studies that evaluated matrix training with individuals with autism spectrum disorder (ASD). Based on 12 reviewed studies, we asked questions regarding participant characteristics, targeted skills, matrix designs, evidence-based teaching strategies, percentage of learning that occurred through direct teaching and recombinative generalization, and the quality of the studies. With 29 total participants, mostly preschool and elementary ages, the most common teaching strategy implemented was a discrete trial training paradigm targeting language development. Overall, findings suggest that matrix training produced an average of 69% of learning without direct teaching across the areas of language, play, sentence construction, and spelling.

A systematic review of the matrix training literature

“…Given the number of procedural decisions that must be made when employing matrix training arrangements, having a means to evaluate the existing relevant research may prove useful. To this end, Curiel, Axe, et al (2020) completed a systematic review of the matrix training literature for individuals with autism spectrum disorder (ASD). They identified and reviewed 12 studies that employed matrix training to teach a variety of skills such as spelling, labeling, and listener identification.…”

Generative time telling in adults with disabilities: A matrix training approach

“…If this does not occur, then the efficiency of matrix training is not employed (i.e., teaching single and two‐component tacts at the same time). Future research could also develop time matrices with the known targets embedded into the initial cells of the matrix and the unknown targets embedded toward the latter portion of the matrix to compare recombinative generalization percentages (Curiel et al, ; Curiel, Sainato, & Goldstein, ; Goldstein et al, ).…”

“…Considering the two‐component features of digital time increments and the numerous teaching targets, matrix training may be an optimal strategy for time telling. Matrix training is a strategy for organizing teaching targets so that only some combinations (i.e., two components responses) are directly taught and subsequent untaught combinations are assessed to determine if they emerge without direct training (Curiel, Axe, Sainato, & Goldstein, ; Esper, ; Goldstein, ). In the matrix training literature, this process has been termed recombinative generalization which is “…differential responding to novel combinations of stimulus components that have been included previously in other stimulus contexts” (Goldstein, , p. 281).…”

“…Matrix training has been shown to produce generative responding for a variety of tact combinations such as noun–verb (Frampton, Wymer, Hansen, & Shillingsburg, ), object–preposition (Pauwels, Ahearn, & Cohen, ), and subject–verb–object (Kohler & Malott, ). A recent matrix training review that focused on individuals with autism spectrum disorder found that there is the potential for learning to occur from direct teaching for approximately 31% of selected target combinations and the remaining 69% to occur through recombinative generalization (Curiel et al, ). Given the positive findings of the aforementioned studies, the nature of hour and minute combinations, and the numerous teaching targets, matrix training is a strategy that should be evaluated for teaching time telling to adults with disabilities.…”

“…Second, this study assessed a diagonal training layout with a matrix containing known and unknown targets. The combination of known and unknown targets was unique because many matrix training studies have used all known or all unknown targets (Curiel et al, ). Both of these options may require extensive pre‐assessments or pretraining.…”