I would ask instead... Are only physiological the important signals to be considered as a source of information about gesture and emotion? If we depart from the physiological (as, for instance micro-electric signals from the body to be transduced into data streams), it does not become more a statistic-oriented work? What about considering a more,"extern" cultural body-manifestations? for instance the sound of laughing? Bio-music should imply, in my opinion, the creation of underlying musical instruments and evolutions based on living organism. And a living organism should be able to trigger gestural and emotional changes in ourselves, not just to provide raw data, to be transduced.  

I think I agree with Carlos. It seems that there is a general push to use the body as a data maker and the computer is the data reader/organizer. In some ways its an over-simplified way to define the human experience. That's to say, to reduce, even the physiological expressions, to a stream of data to be then interpreted non-'living organism[s]' can only ever sound like a list of numbers. The most interesting interactions are those between people (2 or more). The problem is sonification of said interaction. What can we take from Sociology? What can we build into a piece that deals with human expression and which doesn't reduce it to a statistic?  

when I dance I feel like my moves are harmonising with music and recreate it (and mix it a bit too). In fact, my movements are changing my perception of it. Also I do beatbox and i have to say...the human vocals are a beautyfull instrument. Each letter refers to a special sound and you can words for the way they sound. If there was a way i could combine beatboxing with a computer (i wouldnt be surprised if ti already exists) through using the beatbox as a sequencer. Also using pattern recognition to keep the beat going(beatboxing makes you run out of breath and requires some sort of way to keep breathing while keeping the beat and also limits the number of 'tracks'.