RESUMO
Action-concept outcomes are useful targets to identify Parkinson's disease (PD) patients and differentiate between those with and without mild cognitive impairment (PD-MCI, PD-nMCI). Yet, most approaches employ burdensome examiner-dependent tasks, limiting their utility. We introduce a framework capturing action-concept markers automatically in natural speech. Patients from both subgroups and controls retold an action-laden and a non-action-laden text (AT, nAT). In each retelling, we weighed action and non-action concepts through our automated Proximity-to-Reference-Semantic-Field (P-RSF) metric, for analysis via ANCOVAs (controlling for cognitive dysfunction) and support vector machines. Patients were differentiated from controls based on AT (but not nAT) P-RSF scores. The same occurred in PD-nMCI patients. Conversely, PD-MCI patients exhibited reduced P-RSF scores for both texts. Direct discrimination between patient subgroups was not systematic, but it yielded best outcomes via AT scores. Our approach outperformed classifiers based on corpus-derived embeddings. This framework opens scalable avenues to support PD diagnosis and phenotyping.
RESUMO
The linguistic profile of Parkinson's disease (PD) is characterized by difficulties in processing units which denote bodily movements. However, the available evidence has low ecological validity, as it stems from atomistic tasks which are never encountered in real life. Here, we assessed whether such deficits also occur for meanings evoked by context-rich narratives, considering patients with and without mild cognitive impairment (PD-MCI and PD-nMCI, respectively) and matched controls for each group. Participants read two naturalistic stories (an action text and a neutral text) and responded to questions tapping the appraisal of verb-related and circumstantial information. In PD-MCI, impairments in the appraisal of action meanings emerged alongside difficulties in other categories, but they were unique in their independence from general cognitive dysfunction. However, in PD-nMCI, deficits were observed only for action meanings, irrespective of the patients' domain-general skills (executive functions and general cognitive state). Also, using multiple group discriminant function analyses, we found that appraisal of action meanings was the only discourse-level variable that robustly contributed to classifying PD-MCI patients from controls (with an accuracy of 88% for all participants and for each sample separately). Moreover, this variable actually superseded a sensitive executive battery in discriminating between PD-nMCI and controls (with a combined accuracy of 83% for all participants, correctly classifying 79.2% of patients and 87.5% of controls). In sum, action appraisal deficits seem to constitute both a hallmark of naturalistic discourse processing in PD and a sensitive subject-level marker for patients with and without MCI. Such findings highlight the relevance of ecological measures of embodied cognitive functions in the assessment of this population.
