COMPARING TRADITIONAL AND AI-ASSISTED INTERVENTIONS FOR LANGUAGE RECOVERY IN ADULTS WITH APHASIA

Abstract

Background: Aphasia is an acquired language disorder that significantly impairs communication and quality of life among stroke survivors. In Pakistan, the critical shortage of registered speech-language therapists relative to the population creates an urgent need for scalable rehabilitation alternatives.

Objectives: This study compared the effectiveness of traditional speech-language therapy and AI-assisted interventions in improving language recovery among adults with aphasia in Pakistan.

Methods: A quantitative quasi-experimental pre-test/post-test design was employed. Forty adults with mild-to-moderate aphasia were recruited from tertiary care hospitals in Lahore, Rawalpindi and Islamabad, Pakistan, and assigned to either a traditional therapy group (n = 20) or an AI-assisted therapy group (n = 20). Both groups received three sessions per week for eight to ten weeks. Language recovery was assessed across six domains using an Aphasia Language Recovery Assessment Scale. Data were analysed using paired and independent sample t-tests, ANCOVA (controlling for baseline scores), Mann-Whitney U test, Pearson correlation, and Cohen’s d effect size.

Results: Both groups demonstrated statistically significant improvements from pre-test to post-test. However, the AI-assisted group achieved significantly higher post-test scores (M = 78.60, SD = 6.85) compared with the traditional group (M = 68.40, SD = 7.21), yielding a mean difference of 10.20 points, t(38) = −4.58, p < .001, 95% CI [−14.76, −5.64]. Cohen’s d values ranged from 0.94 to 1.45, all indicating large effects. ANCOVA confirmed the group difference after controlling for baseline scores, F(1, 37) = 19.61, p < .001, partial η² = .344. Pearson correlation revealed a moderate negative association between diagnosis duration and recovery scores (r = −0.48, p = .002).

Conclusion: AI-assisted intervention produced superior language recovery outcomes across all domains compared with traditional therapy. Given Pakistan’s severe speech-language therapy workforce shortage, AI-assisted rehabilitation represents a scalable, cost-effective, and linguistically adaptable solution for aphasia rehabilitation. Early referral and integration of AI tools into national rehabilitation policy are strongly recommended.