Wednesday, 6 August 2014

cellF #2 dish of neurons control NLC modular


  1. Andrew can you explain this a bit more.
    Looks very interesting

  2. We have a bunch of neurons growing in a dish. the dish has about 60 electrodes to detect electrical signals from the neurons and to zap them with signals (about 1V p-p). The neurons give off a constant chatter called 'neuronal noise', something like white noise but it seems a lot easier on the ear.....maybe more organic ;)
    Some of the neurons give off spikes called 'action potentials' these occur in various ways, sometimes erratically and rarely, sometimes with patterns and frequently.
    A third behaviour, known as 'local field potential', occurs when the neurons give something like a great big group hiccup en masse, these are rarer so everyone gets quite excited to see them. You can see a couple in the vid.
    I made a circuit to amplify the neuronal noise to synth levels, plus a comparator to detect the action potentials and turn them into synth level triggers. Another circuit converts synth signals into triggers at a voltage and shape pleasing to the neurons (large falling edge).
    In the vid, you can hear the neuronal noise filtered thru a Moog vox filter clone, also the neuronal noise is patched into the CV input of a VCO and we are also listening to the pulse from that via a VCA
    The triggers are used to clock the sequencer and ADSR, the pulse divider on the sequencer is sending back a trigger to the neurons every 8th to 128th time the neurons send a trigger out. They seem to like this feedback, without it they tend to get bored and inactive.
    On the computer, you can see each cell is recording the activity being detected on each electrode. It is useful for us to see this to decide which electrodes to patch into. Eventually this will be done with a LED array. despite the computer the connections between the neurons and the synth are all analogue.