Our brains are amazing predicting machines.
In fact, even if you were relaxing watching your favourite DVD on mute and not even paying attention to tones that are playing in the background, your brain is continually predicting what it will hear next. In what’s referred to as an Oddball paradigm, you can present a series of tones at regular intervals and then throw in an ‘oddball’; a one-off different tone (either higher / lower, louder / softer, sooner / later it doesn’t seem to matter in which way its different). When you compare the brain’s response to this deviant tone with the standard tone using MEG or EEG, researchers find the “Mismatch Negativity” (MMN), as soon as 90 – 250ms after the tone, which is basically the brain’s response to this prediction error. Though at such a short timescale, this is your brain detecting this error before you even realise.
Recently Claudia Lappe and colleagues at University of Muenster wanted to untangle this prediction error further to tease apart whether there are differences between not just prediction errors in pitch or tempo (as described above but the oddball tones were either lower in pitch or came in earlier), but also if differences in prediction errors between melodies and rhythms could be seen. This is a crucial undertaking because most of the music we listen to is not a stream of a single note – it’s within the context of a piece of music, with both melody and rhythms interacting.
Firstly Lappe et al looked at when and how large was the average whole brain’s response to these four types of unexpected occurrences (called “deviants” in this figure). They found that within a musical context a note out of time (last note coming in too early) showed a significantly larger MMN, and was also significantly sooner than the melodic violation (last note out of chord) – even though the timings of the onset of the tone were re-aligned for analysis. However, no difference in the timings of the MMN were found in the pitch and tempo simple tone pattern errors, though an oddball different pitch showed a larger MMN response than an oddball shorter duration note.
So averaging over the whole brain clearly there are differences between violating expectations in melody or rhythm, but how are these generated in the brain? Is it the same parts of the brain that fire differently, or if there are actually different regions involved in detecting these different errors? To do this Lappe et al used a type of beamforming analysis.
They found that melodic and rhythmic deviations both activated auditory regions and the supplementary motor area (interestingly both known to be involved in musical imagery). However melodic deviations also activated frontal brain regions, whereas rhythmic deviations activated parietal regions. The pitch and tempo violations in simple patterns were really just limited to auditory regions.
These results aligns with theories that suggest what we are hearing is processed in a parallel but different stream to when or where spatially we are hearing it. We are processing what we hear (even passively) in these two streams, and continually predicting what we will hear next. And hence the violations to our expectations in timing (when) are being processed faster than violations in expected melody (what). This also confirms previous studies of Lappe et al where non-musicians underwent short term piano training; the “what” was processed more slowly than the “when”.
This is an exciting paper because of the questions that it raises for further investigations. For example, if a wrong note is both early and out of key, how would the MMN compare – could the two pathways (what and when) both be seen in the response, and would the “when” response still be quicker? What if the wrong note was late (in which case the MMN would be measured from the omission of the expected tone rather than the presence of it)? And how does all this effect the way we play music together? These are just some of the questions that Lappe is continuing to investigate.
For more information about the concept behind a #3MinutePaper check out this blog.