Behavioural data analysis results

[skjerns]

I've added the behavioural results

It seems like the manipulation for valence worked very well, not that good for arousal. There are some participants for which the correlation of valence is very low, which means that they rated negative GIFs positive and vice versa. Might be an idea to exclude these participants or pay extra attention to them (eg ERP19). Also they rated low arousal GIFs as high arousal and vice versa (e.g. ERP05)

Same shown in this plot, valence ratings nicely separate between positive and negative. For arousal, it seems like negative GIFs show much higher arousal. But in general the arousal of mean 3+ is quite okay I would say, so we succeeded in eliciting arousal in participants.

[skjerns]

Button presses

Buttons were more often pressed for negative emotions. This is potentially important, as the button presses should be decodable in the MEG, e.g. we could potentially just decode a motor response if we include all trials and have as a target valence.

per participant we see that most participants are quite consistent and either press or do not press the button (e.g. ERP19 almost never pressed the button)

Some participants have bias and press more on negative than on positive buttons. However the effect is not massive

Buttons are also more often pressed when subjective arousal is high and when valence is either very high or very low. That is great!! What we would expect.