External location of touch is constructed post-hoc based on limb choice

Femke Maij; Christian Seegelke; W Pieter Medendorp; Tobias Heed

Having established that participants appear to have perceived the stimulus 259 at the reporting hand, two possibilities remain as to which stimulus location was 260 associated with erroneous responses (see Fig. 1 ). The stimulus switch hypothesis 261 posits that the two stimuli were localized correctly, and one is chosen for the 262 response. In TOJ error trials, participants would confuse the two stimuli and report 263 the second stimulus by pointing at its location with the respective, incorrect hand. The copyright holder for this preprint (which was not peer-reviewed) is the author/funder. It . https://doi.org/10.1101/549832 doi: bioRxiv preprint curves. Such movement time-related biases were evident also in the present data 295 (see Fig. 5 ). In the case of correct TOJ trials, we assume that participants aimed, 296 as instructed, at the location of the correct hand at time 1. Therefore, we derive the 297 localization error curve as the spatial difference of reaching endpoint and hand 298 location at time 1 (see Fig. 6 , dark blue lines). However, we can also derive a 299 . CC-BY-NC-ND 4.0 International license is made available under a The copyright holder for this preprint (which was not peer-reviewed) is the author/funder. It . https://doi.org/10.1101/549832 doi: bioRxiv preprint The copyright holder for this preprint (which was not peer-reviewed) is the author/funder. It . https://doi.org/10.1101/549832 doi: bioRxiv preprint four start and end posture combinations. To quantify this further, we computed by 317 how much reaching endpoints of TOJ error trials had to be shifted to best match 318 either template curve derived from correct TOJ trials. If, in incorrect TOJ trials, 319 participants aimed for the incorrect hand's position at time 1, then the temporal 320 shift should be zero relative to the localization error curve in correct TOJ trials 321 relative to time 1; furthermore, it should be about -110 ms (negative denoting a 322 shift towards left, see above) compared to the localization error curve of correct 323 TOJ trials relative to time 2. If, however, participants aimed for the incorrect hand's 324 position at time 2, the shift pattern should be exactly reverse, that is, zero 325 compared to the second template curve, and near +110 ms compared to the first 326 template curve. The copyright holder for this preprint (which was not peer-reviewed) is the author/funder. It . https://doi.org/10.1101/549832 doi: bioRxiv preprint Point ( 2 (9,10) = 15.55, p < 0.001), but no effect of Posture Condition ( 2 (7,10) = 345 1.57, p = 0.67) or interaction ( 2 (7,10) = 4.53, p = 0.21) between the two factors. The copyright holder for this preprint (which was not peer-reviewed) is the author/funder. It . https://doi.org/10.1101/549832 doi: bioRxiv preprint marks the computation of tactile stimulus location based on correct stimulus timing 505 and movement information of a (correctly or incorrectly) implied body part. 506 Accordingly, limb crossing affected hand assignment, but not stimulus localization. 507