Componential Network for the Recognition of Tool-Associated Actions: Evidence from Voxel-based Lesion-Symptom Mapping in Acute Stroke Patients

The study aimed to elucidate areas involved in recognizing tool-associated actions, and to characterize the relationship between recognition and active performance of tool use. We performed voxel-based lesion-symptom mapping in a prospective cohort of 98 acute left-hemisphere ischemic stroke patients (68 male, age mean +/- standard deviation, 65 +/- 13 years; examination 4.4 +/- 2 days post-stroke). In a video-based test, patients distinguished correct tool-related actions from actions with spatio-temporal (incorrect grip, kinematics, or tool orientation) or conceptual errors (incorrect tool-recipient matching, e.g., spreading jam on toast with a paintbrush). Moreover, spatio-temporal and conceptual errors were determined during actual tool use. Deficient spatio-temporal error discrimination followed lesions within a dorsal network in which the inferior parietal lobule (IPL) and the lateral temporal cortex (sLTC) were specifically relevant for assessing functional hand postures and kinematics, respectively. Conversely, impaired recognition of conceptual errors resulted from damage to ventral stream regions including anterior temporal lobe. Furthermore, LTC and IPL lesions impacted differently on action recognition and active tool use, respectively. In summary, recognition of tool-associated actions relies on a componential network. Our study particularly highlights the dissociable roles of LTC and IPL for the recognition of action kinematics and functional hand postures, respectively.