You are viewing the SNL 2017 Archive Website. For the latest information, see the Current Website.

Poster A3, Wednesday, November 8, 10:30 – 11:45 am, Harborview and Loch Raven Ballrooms

Reduced Stroop competition between tool action “neighbors” in left hemisphere stroke

Harrison Stoll1, Tamer Soliman1, Laurel Buxbaum1;1Moss Rehabilitation Research Institute

Producing a word or action requires selection from among a set of competing alternatives. Response competition strongly affects action errors in left hemisphere stroke patients with limb apraxia. Apraxics pantomiming the use of tools make more errors with tools associated with conflicting actions for use versus grasp-to-pick-up (e.g., TV remote; i.e. “conflict” tools) than tools having a single action for both use and grasp (e.g., hammer; i.e. “non-conflict” tools) (Watson & Buxbaum, 2015). This susceptibility to competition from grasp actions is associated with weakness of ‘automatic’ activation of use actions, and reduced competition from use ‘neighbors’ that share sensory-motor attributes such as hand posture (Lee et al., 2014). A predicted consequence of weakened automatic activation of use representations and reduced competition between action ‘neighbors’ is relatively good performance when neighbors are directly pitted against one another in a Stroop-like task requiring inhibition of a typically pre-potent use action. We tested this hypothesis in a group of 21 left hemisphere stroke patients. In each of 6 blocks, a picture of one of 2 tools was shown over 24 trials. Within each block, the 2 tools were either Action neighbors (e.g., key and lightbulb), Function neighbors (e.g., axe and saw), or Unrelated. On Congruent trials in each block, subjects produced the action appropriate to the tool. On Incongruent trials they produced the action appropriate to the other tool in the block. The Stroop cost was relative performance on Incongruent versus Congruent trials. Patients also performed a simple pantomime task requiring them to gesture the use of both conflict and non-conflict tools. We computed overall scores and a use conflict score (relative performance with conflict versus non-conflict objects) -- an index of use activation strength. Conflict scores were uncorrelated with overall scores (p > .1). Data were analyzed with mixed-effects models with condition (Action-related, Function-related, Unrelated) and Congruence as within-subjects factors. On the action Stroop task, Incongruent actions were less accurate than Congruent actions. More interestingly, higher effects of conflict on the simple pantomime task were associated with lower Stroop costs in the Action condition (p = < .05), but not in the Function or Unrelated conditions (p’s > .5) (see Figure 1). Moreover, the relationship between the pantomime conflict effect and Stroop costs was stronger in the Action condition than in the other two conditions (p’s < .01). Consistent with our hypothesis, reduced activation of use information in a pantomime task is specifically associated with reduced Stroop costs from close action neighbors, but not from close function neighbors. There are at least two possible accounts of these data. Less precise activation of tool use actions may lead to spreading activation in the action “neighborhood”, priming close Incongruent actions and making them easier to produce. Alternatively, weakened automatic activation of use actions may be associated with weaker “lateral” inhibition of neighboring (in this case, Incongruent) actions, rendering them easier to produce. Computational modeling of the observed effects may be useful in disentangling these possibilities.

Topic Area: Control, Selection, and Executive Processes

Back to Poster Schedule