For the last two weeks I've been trying to develop an operant learning paradigm for isolated leech nervous systems. To do these experiments I traveled to visit Bill Kristan's lab at UCSD. After he showed me how to dissect the animal, I went and did some pilot studies on where to stimulate the animals and how. The plan was to first get the animals to generate swimming or crawling behavior in response to stimulation of a nerve at ganglion 15, just as described in Kevin Briggman's Science paper. The initial problem was that the animals all swam spontaneously, as shown in this video:



This spontaneous swimming behavior makes it impossible to decide which activity was spontaneous and which was triggered by the stimulation. Moreover, it seemed that this spontaneous swimming was an indication that the preparations were very prone to respond with swimming to any kind of stimulation, anywhere. I think I have the problem of spontaneous swimming solved now by removing some of the connective tissue around the ganglia (it's actually a blood sinus called the 'stocking'). This should prevent the accumulation of serotonin in the blood sinus, which biases the system to swimming activity. I also had to move the stimulation from the published ganglion 15 down to ganglion 17 or even 18 in order to get any crawling at all (apparently some weird seasonal thing). Still, it doesn't seem I'm getting a 50-50 distribution of swimming vs. crawling, yet. I need to play around with the duration and intensity of the stimulation it seems. The last prep today almost exclusively crawled, sigh. Anyway, I seem to be making headway.

Progress is slow, though. The dissection takes about 3 hours, then the nervous system gets 1h of rest (aka lunch) and then the trying out, testing, recording and stimulating takes about another 3-4 hours, which makes it too long to get two decent preps into a single day. I need to somehow shave off a few hours from this schedule, I only have three weeks left.

Why am I doing all this, you ask? Well, if the nervous system is making a decision upon each stimulation whether it should generate crawling or swimming activity, maybe I can find a nerve to stimulate which would bias this decision one way or another. So I've started to stimulate a nerve in the very front of the nervous system, ganglion 2 or 3. When I do that during a swim I can usually switch it off. If I do so during a crawl, they usually stop crawling, too. However, the method is not as reliable as I would want it, yet. The idea behind this is that eventually I should be able to get preps where in a pre-test phase they exhibit (ideally) a 50-50 ratio between swimming and crawling in response to the stimulation at ganglion 17/18, Then, during training, I choose one of the patterns and shut them off with the ganglion 2/3 stimulation. After training, I hope to see a shift in the ratio, away from the initial one, becasue the nervous system has learned that the pattern in question is always followed by some unpleasant stimulation.

I've done similar operant conditioning experiments in the marine snail Aplysia, so these kinds of things can work in principle. But which nerve to use and which stimulation parameters is always tricky. It may well be that I can't get the nerve chords to modulate their decision-making.