Traditional clinical and research perspectives have often attributed post-stroke pragmatic deficits—that is, difficulty interpreting communicative intentions and non-literal meanings—primarily to damage of the right hemisphere. Left-hemisphere damage (LHD) has typically been linked to aphasia and structural language impairment. This study set out to examine whether pragmatic impairment after stroke depends on lesion side by assessing pragmatic profiles across LHD and RHD patients and comparing them with healthy controls.
The sample included 99 stroke patients: 40 with left-hemisphere damage (including 14 with aphasia of minimal-to-moderate severity) and 59 with right-hemisphere damage, plus 60 healthy control participants. Pragmatic abilities and related cognitive substrates were measured with the Assessment of Pragmatic Abilities and Cognitive Substrates (APACS), which yields task-level scores and composite indices such as APACS Production and APACS Total.
The authors used three converging, complementary analytic strategies to probe hemispheric differences in pragmatic profiles:
Associations between pragmatic performance and cognitive or language measures were also evaluated, focusing on variables such as Theory of Mind, structural language abilities, and general cognition.
Across the APACS battery, stroke patients as a whole performed worse than healthy controls, indicating that pragmatic abilities are vulnerable after stroke irrespective of lesion side. This established a clear group-level difference between patients and non-brain-injured participants.
Permutation testing revealed largely no significant differences between LHD and RHD patients on most APACS tasks. Two exceptions emerged: tasks that required expressive verbal output—the Interview task and Figurative Language 2—showed worse performance in the LHD group. These task-level differences contributed to lower APACS Production and APACS Total composite scores for LHD patients.
Equivalence tests supported the conclusion that most pragmatic measures were statistically equivalent between lesion sides. The same expressive tasks and composites flagged by permutation testing (Interview, Figurative Language 2, APACS Production, and APACS Total) did not reach equivalence and therefore remained inconclusive rather than showing robust hemispheric divergence.
Unsupervised clustering of patient profiles did not sort patients according to lesion side. In other words, clusters of pragmatic impairment patterns contained both LHD and RHD patients, providing further evidence that lesion laterality was not the primary organizing factor for pragmatic deficit profiles in this dataset.
Across both lesion groups, Theory of Mind emerged as a strong correlate of pragmatic performance, indicating the importance of social-cognitive processes for pragmatic ability after stroke. Differential associations were observed for other substrates: structural language measures related specifically to pragmatic performance in LHD patients, whereas measures of general cognition showed associations with pragmatic performance in the RHD group.
When the authors excluded the subset of LHD patients with aphasia, evidence for hemispheric differences weakened further. This suggests that the residual LHD disadvantage on expressive tasks was largely driven by aphasic patients within the LHD subgroup rather than reflecting a general left-hemisphere specialization for pragmatics.
The principal conclusion is that primary pragmatic impairment—particularly in receptive components—occurs comparably after left- and right-hemisphere stroke. The only clear residual disadvantage for LHD was confined to tasks demanding open verbal output, and this appeared to be driven in part by co-occurring aphasia. These findings challenge the traditional assumption of right-hemispheric specialization for pragmatics and support the practice of assessing pragmatic abilities in all post-stroke patients regardless of lesion laterality.
This report is a preprint and has not undergone peer review. The article indicates the presence of supplementary materials and reports specific sample counts and the analytic approaches used, but further methodological details, full statistical tables, and any additional demographic or lesion-mapping information are available in the full preprint and supplementary files. The authors declared no competing interests. Funders are listed in the source.