3 ms·
I'm looking at a dataset now of people listening to 12 music tracks with 1000hz sampling rate 128 channel EEG. We are trying to predict which tracks the 25 peop
by dr_dshiv 7y ago
I'm looking at a dataset now of people listening to 12 music tracks with 1000hz sampling rate 128 channel EEG. We are trying to predict which tracks the 25 people are listening to, based on the EEG.
It's a good ML problem. We have accuracy up to 85% when predicting person&track together, falling to 30% when just predicting track.
Since we also had people rate enjoyment and familiarity, the main theoretical question is whether there is an direct neural resonance with the music during highly enjoyed tracks. I.e., that more of the brain is coupled to the music when you feel it deeply.
- kranner 7y agoIs this a public dataset?
- ghastmaster 7y agoYou made me consider whether EEG would notice differences when listening to 24bit/192hz vs. 16bit/44.1hz audio. So far the articles I have read only pertain to the logical science and or anecdotal individual double blind tests. Thank you.