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

 
Poster B11, Wednesday, November 8, 3:00 – 4:15 pm, Harborview and Loch Raven Ballrooms

Tracking keystroke sequences at the cortical level

Svetlana Pinet1,2, Gary S. Dell3, F.-Xavier Alario2;1Johns Hopkins University, 2Aix-Marseille Universite & CNRS, 3University of Illinois at Urbana-Champaign

Language production must solve the problem of serial order when all items of a sequence (e.g., letters of a word) get activated but each should be launched at the appropriate time. Several proposals postulate the existence of inhibitory processes to achieve correct ordering (Houghton, 1990; Rumelhart & Norman, 1982). However, the cognitive and physiological implementations of this inhibition remain to be determined. In this study, we recorded EEG during typing. Our previous work consistently reported a meaningful pattern of electrophysiological activities over motor cortices prior to the first keystroke of a sequence, showing that the activation of the contralateral motor cortex co-occurs with the inhibition of its ipsilateral counterpart (Pinet et al., 2015, 2016; Scaltritti et al., 2017). Because sequences provide contrasting environments in terms of recruited effectors (e.g., bimanual vs. unimanual), we expected activation and inhibition processes to develop differently according to sequence properties. Eighteen right-handed participants were selected for their ability to type without need for looking at their hands. EEG was acquired via 128-channels while participants typed short words on a computer keyboard. Stimuli (both auditory and visual) were divided into three conditions: single (one letter), unimanual (two-letter words, typed with one hand), bimanual (two-letter words, typed with two hands). Over the contralateral hemisphere, all conditions presented a negative (activation) component before the first keystroke and unimanual sequences were associated with higher negative amplitude. Over the ipsilateral hemisphere, both single and unimanual conditions presented a positive (inhibition) component, while bimanual sequences exhibited a negative component. Statistical analyses (cluster-based non parametric analyses) confirmed these differences. In the case of bimanual sequences, both hemispheres exhibited negative components with similar time-course, albeit shifted in time. A detailed analysis of the specific time courses allowed the conclusion that the activations observed over each hemisphere reflected the temporal sequence of the two hands required for the response. These results show that the activity recorded over motor cortices prior to sequence execution is finely tight to the content of the prepared sequence. Both contralateral and ipsilateral components are modulated by different factors, which points to their relative independence. Because differences were already observed before the first keystroke, sequences appear to be substantially prepared before execution begins, compatible with keystroke pre-activation. We also observed similar activation of a single keystroke, whether isolated (single letter) or part of a (bimanual) sequence, which argues for unitary activation processes. Moreover, a positive inhibitory component was only observed in unimanual contexts, regardless of the length of the sequence. Thus, inhibition appears tied to some characteristics of the sequence produced, such as the effectors involved, but independent of the number of elements. This strongly suggests the existence of a processing step dealing with selecting involved effectors that will then deploy a flexible system for inhibiting alternate effectors if necessary. Although vastly ignored in previous cognitive models of sequence production, we showed that the effectors recruited in a sequence constitute an important processing level. Our results could help bridge the gap between cognitive hypotheses and neurophysiological implementation of inhibition.

Topic Area: Writing and Spelling

Back to Poster Schedule