gms | German Medical Science

Objective: Response latency, in addition to response accuracy, provides a very powerful measure in psycholinguistic neurostimulation approaches and could potentially also be used to identify language eloquent brain regions in TMS guided language mapping. However, objective response latency measures are rarely seen in language mapping paradigms, as so far these latencies could only be acquired post-hoc involving significant effort. A solution was presented by Vitikainen and coworkers (2015) who introduced the Accelerometer as a convenient device to automatically detect vocal response onsets on the basis of larynx vibrations in language mapping paradigms. The current approach aimed at validating the temporal precision of Accelerometer data.

Methods: A word reading paradigm was conducted on a cohort of healthy participants, while an Accelerometer was placed on their throat to measure larynx movements during speaking. Word stimuli of the reading paradigm were selected to cover a wide range of word-initial phonemes and each word was presented three times. Automatically detected Accelerometer response latencies were validated by comparison to manually determined responses from sound recordings. Effects of word-initial phonemes and repetitions were investigated with repeated measures analyses of variance.

Results: Results indicate an overall good match between manual and automatic detection of response latencies. Critically though, systematic differences in Accelerometer precision between word-initial phonemes were observed (p<0.001). At the same time, word-initial phoneme specific precision of the Accelerometer was shown to not vary significantly over time (p=0.5).

Conclusion: Accelerometers may very well be used to automatically detected response latencies in TMS guided language mappings. Results must however always be related to word-specific baseline performance to account for word-initial phoneme depended differences in Accelerometer precision.