I'm quite new to mediation analysis so hopefully I can clearly explain my problem.
I wish to conduct a mediation model with a single mediating variable. The independent variable is a time series EEG data, the mediator variable is a time series EMG data, and the dependent variable is a single data point (emotion felt during the test given as intensity given after the recording). The EEG and EMG was recorded simultaneously during the study whilst the participant was exposed to an emotional stimuli. The emotion felt was then reported by the subject after the emotional stimuli ended. I wished to find the possible effect the muscles recorded with EMG on the EEG recordings. Whether causal or correlational.
Is there any way to keep the independent and mediator variable in the form of a time series if I were to apply a mediation model? Or would I have to convert the EMG and EEG data into single data points? Would you suggest a different model that would keep the data in it's original form?