More complex behaviour in humans and animals tends to be organised (and learned) in a series of steps, a sequence of actions in time that allows the brain to control better the full event. Think about when you try to calculate a large sum, or vow someone you fancy or even when learning kung-fu fighting. You go slow, by parts, and, as you get a reward (or not), positive (or negative) feedback, you move to the next step. In these “self-learned” (also called voluntary) tasks the signals to start and stop the action are crucial to achieve the proper behaviour and get the result you want. Obsessions are, exactly, an incapacity to know when to stop.
But how does this process work within the brain? Nobody really knows.
But scientists suspect that the substancia nigra and the striatum (two structures in the basal ganglia) are somehow associated to this learning process. In fact, PD patients – which have learning and motor coordination problems (thus the characteristic tremors) – and addicted or obsessive individuals have all been shown to have abnormalities in the striatum and in the nigrostriatal circuit. The nigrostriatal circuit is the nerve fibres that pass information from the substancia nigra to the striatum using dopamine – the “pleasure” (or reward) neurotransmitter of the brain. In fact, scientists also know that dopamine is crucial for learning. After all when something feels good we want to go back for more.
But despite all these clues and the importance of the diseases associated with problems in self-learned behaviours not much had been done to understand better the neural and molecular mechanisms behind them. In an attempt to change this Xin Jin and Rui M. Costa at the Laboratory for Integrative Neuroscience, in National Institute on Alcohol Abuse and Alcoholism, Maryland, USA and the Champalimaud Neuroscience Program at Instituto Gulbenkian de Ciência, Portugal decided to look into this problem in mice, which are a particular good subjects for this type of experiments as not only their brain is very similar to ours, but also their genetic is well known and they are easy to train and analyse.
In Jin and Costa’s experiments the mice were trained to press a lever 8 times in order to receive a sugar reward, while neural activity in their striatum and substancia nigra was recorded. This is a typical selflearned task as the mice through trial and error teach themselves the right number of presses needed to get the reward it wants.
It was found that – as training progressed and mice learned the task – two major things happened: the animals stepped faster on the lever, taking less and less time to press it 8 times (so become more confident and efficient), and the numbers of steps becomes less variable until they are discreet series of 8. Optimal behaviour (so with little variability in their stepping pace or number of steps while finishing the task quicker) was achieved after 6 days, the time usually necessary to train these animals.
Measurements of the neural activity in
the striatum and substancia nigra of these
mice revealed that neurons in the striatum
and the nigrostriatal circuits had a particularly
unique increase (a spike) of electrical
activity just before the start and the end of
each set of 8 steps (so before step 1 and
step 8). With training – and while the
number of activated neurons stayed the
same – the peaks of electrical activity
before step 1 and 8 in these areas became higher – showing a better signalling,
almost like turning on a (brighter) green or
a red light to start or to stop the behaviour.
Interestingly, most neurons only signalled the start OR the stop signal but not both, and the ”spikes” were specific for each action as response to different levers induced different levels of neural activity. Both these characteristics are important as they show that these neural responses are very specific which also means that targeting individual neurons can achieve extremely fine tuned treatments.
To confirm these results Jin and Costa next created mice incapable of receiving the signals of the nigrostriatal circuits (by disrupting a molecule involved in recognition of dopamine in the striatum – no dopamine, no neural messages) and put them through the same protocol to find that these mice, contrary to those with an intact circuit, were not able to self-learn.
When looking at their brain the researchers found that they had much less start/stop neurons than normal mice, but also that the electric activity (the spikes) of the neurons remaining did not increase with time. And although mice were able to understand that pressing a lever would give them sugar, they were not able to learn to become more efficient getting it. In fact, neither the time it took them to press the lever 8 times become shorter (so more efficient) neither they ever managed constant groups of 8 steps like normal mice, instead their number of presses varied all the time.
From these experiments Jin and Costa were able to conclude that dopamineproducing neurons coding stop/end neural signals in the nigrostriatal area were crucial to a capacity to learn voluntary movements (like kung fu fighting…).
This is an important result because as we understand better the molecular basis of neural circuits the more specifically and effectively we can treat patients, and selflearned behaviours (and disorders) are of extreme importance in the life of cognitive animals, particularly humans.
But the discovery has also very direct implications – for example in PD – a disease where patients lose the capability to learn and control voluntary movements and are known to have problems in the striatum function. At the moment – because these patients lose dopamineproducing neurons – the disease is treated with drugs that increase the levels of this neurotransmitter in the brain. But what Jin and Costa’s experiments seem to suggest is that it is not so much the general dopamine levels that are important here, but, instead, the capacity of a few very specific (start/stop) neurons to produce boosts of the neurotransmitter when necessary. This, if confirmed, would imply the need of a very different therapeutic approach to PD.
But Jin and Costa’s results also have implications for obsessive and compulsive behaviours, which is now believed to be the result an aberrant self-reinforced positive feedback – the more they do something, the more this feedback tells them to do it. “A possibility after these results,” says Rui Costa, “is that the problem in these individuals is not simply on the positive feedback, but that they also lack a proper negative feedback – maybe their stop neurons are affected and they simply cannot quit. This actually something we want to investigate in the future.”
Another potential implication is in the fields of learning and behavioural therapies used in these diseases. These can be used even in PD where, for example, we now know that patients react better to negative feedback – they even learn better than normal patients when this is used – due to the fact the neurons stimulated in these conditions use very little dopamine. Approaches that rely on reward simply do not work because PD patients do not have enough dopamine. In the same way patients identified with problems in the stop or start neurons can, in the future, receive behavioural therapy developed to overcome this particular problem.
There are many potential applications that one day can be developed from having discovered these remarkable neurons but for now, as Rui Costa points out “the plan is to identify better these start and stop cells and see if we can manipulate their activity”.
● Reference: Start/stop signals emerge in nigrostriatal circuits during sequence learning, Xin Jin & Rui M. Costa, Nature Vol. 466, doi:10.1038/nature09263 www.nature.com/nature/journal/v466/n7305/abs/nature09263.html
Copyright © 2011 MiddleEastHealthMag.com. All Rights Reserved.