I was attending the Science of Learning Symposium at JHU, which is usually quite fascinating, but was very boring this year. These are the things that frustrate me about formal events:

1. People take extended time to introduce the speakers (nobody wants to hear you talk, go away).
2. People in suits, usually senior professors who have stopped being scientists, who think they’re funny because nobody is honest to them, make shitty jokes and laugh at their own jokes.
3. People make a huge preamble when asking questions, dude the question time is very limited and there are several others who also have things to ask.

It is at formal events like these that I sympathize strongly with Salinger’s protagonist from Catcher in the Rye.

But then this speaker came to the front and before starting his talk reminded everybody in the audience, how privileged they were to be able to attend a university and all that, and that it is everybody’s responsibility, more than ever today, to contribute back to the society. He quoted this quote by Einstein:

“The world is a dangerous place, not because of those who do evil, but because of those who look on and do nothing.”

And I knew right away that this was a talk worth listening to.

David Poeppel talked about his work on the recognition of spoken language in the brain, about how the brain converts the pressure fluctuations in the air, as detected by the ear, into language. There were several great insights in his talk. One of the biggest differences between language processing when we hear as opposed to when we read, is that in spoken language there are no whitespaces between words, so our brain has to constantly figure out where the whitespaces are.

Following were some of the things I learnt at this talk:

1. Our frequency of speaking is almost constant, in fact, the frequency of syllables is almost the same across languages (I think this is an oversimplification as surely some people speak faster than the rest).
2. Our brain activity oscillates, as detected by an MEG.
3. When we hear someone speak our brain activity oscillates at the same frequency as that of the syllables we’re hearing!
4. When we speak we also have a fixed (lower) frequency with which we speak the complicated structures like phrases, and sentences (again this sounds like an oversimplification). This frequency is what allows the brain to determine which syllables group together to make a word.
5. The brain activity also oscillates at these lower frequencies i.e. if you did a Fourier transform of the brain activity you get peaks at syllable frequency, phrase frequency, and sentence frequency!!
6. This synchronization of frequency needs active focus, if there are multiple people who’re speaking in the background with different frequencies their frequencies do not show up.
7. This is not a statistical phenomenon. Not sure I understand what this means, but something like the brain is not learning dynamically from the sound but rather the language processing units are involved in determining the frequencies.
8. By seeing which parts oscillate at what frequencies, it is in fact possible to detect which parts of the brain are needed for processing syllables, phrases, and sentences!!!

References: Slide-1, Slide-2.