Between August and October 2017, I spent 6 weeks in Marseille, at the Institut de Neurosciences de la Timone (INT). The main reason to go there was because of their excellent knowledge on EEG analyses, and of their work with the KINARM, a robotic device used for force-field and spatial perturbation tasks. During my stay I was supervised by Nicole Malfait and I got to learn more about the project done by Amirhossein Jahani (ESR-13). Amir’s project is closely related to my own, as we both study error correction in human motor learning. The benefit of doing my secondment at INT was that, while I usually only focus on the behavioral processes of motor control, I got to learn more about the neural processes involved.
The objectives of the project were to learn to collect EEG and analyze it with Matlab and Fieldtrip. I specifically got to learn to do Independent Component Analyses (ICA) on EEG data. The next step was to collect and to analyze some of the data for the main project. The project focused on the modulation of EEG β-band in medial post-central areas of the brain. A study of Tan et al. (2014) showed that the β-activity (an increased activity in the β-band frequencies at the end of a movement) increases with decreasing error size. In a study of Torrecillos et al. (2015) it was found that the fore-period was also modulated by error-size. This project focused on disentangling implicit sensorimotor adaptation processes, and explicit mismatch processes. The main idea is that foreperiod β-activity might be more related of the implicit processes of updating sensorimotor mismatches, while post-movement β-activity might reflect explicit mismatch/surprise when experiencing an error. For that reason an experiment in which subjects learned a rotational perturbation in short learning-blocks was designed. Similar to the experiment of Mazzoni & Krakauer (2006), subjects were exposed to trials of rotation during which they explicitly changed their aiming-strategy. The perturbation blocks were only 4 trials long, and subjects were informed when the perturbation was not present anymore. Subjects could account for errors during the perturbation blocks (ROT1-3), but started over-adapting, as was shown before by Mazzoni & Krakauer (2006). The first trial after the perturbation (after effect; AF trial) block should therefore contain the implicit strategy that was learned during the perturbation block. If the modulation of foreperiod β-activity in the sensorimotor areas is indeed involved in the implicit learning of sensorimotor areas, this should be visible during the AF trials, as the explicit strategy is not present during these trials, but the implicit learning is. Below I show some (EEG-ICA) results for one example participant during a rotation (ROT) trial compared to the baseline (BL). We can see that the ROT trials, in which the strategy is applied, the β-activity is modulated. For this participant we could not find the same for the AF vs BL, but a decreased activity in AF trials vs BL trials was found when group analysis was done.
For the full extend of the project I’d like to refer you to the page of Amir (itn-pace.eu/amirhossein-jahani/).
During the 6 weeks I took part in some data-collection for this project, and also was a participant myself. I analyzed the data of some of the participants. Thanks to Nicole Malfait and Amirhossein Jahani I have learned a lot about EEG methods, ICA and the neural processes involved in sensorimotor learning. My stay at INT Marseille also taught me about the workings at other labs, which was very insightful. I very much enjoyed this secondment.