negative feedback
21 August, 2011
Filed in Category: horse behaviour
Negative and Positive FeedBack Control Systems
Many of you already know that I have an academic background in physiology. During my career in medical research I was required to teach a handful of courses to postgraduate science students as well as medical and veterinary students. One of the subjects I lectured on was control systems in physiology ie, how does the body keep itself working within normal parameters. By that I mean how does the body stay within a normal temperature range and regulate itself when the temperature is too high or too low. The same for heart rate, oxygen levels, blood pressure, appetite, body fluid etc – virtually every function in the body is regulated by a control system that turns it on or turns it off.
Most control systems use negative feedback loops. For example, if your temperature gets too high your body sees this and activates functions like sweating and increasing blood flow to the skin in order to lose heat and cool the body down until the brain decides your body is back in a normal temperature range. Nearly all the body functions use a similar type of control system to keep them within a normal range.
But a small number of the bodies functions don’t use negative feedback loops, some use positive feedback. Positive feedback systems are potentially dangerous because the only limitation on their function is when a catastrophic event occurs. In a positive feedback system, one event leads to a slightly bigger event, which then leads to an even bigger event and this leads to a much bigger event. It’s like there is a cascade of events and causes one causes even more to happen, rather than less (as occurs with a negative feedback system).
When I was teaching, in order to wake the students up about this boring topic I would ask them, “As physiologists, what is the most exciting thing about an orgasm?” That would get their attention. But of course the answer is that it is one of the few examples of a positive feedback loop. It starts slow and accelerates into a system almost out of control until orgasm occurs (the catastrophic event). Another example is birth where it begins with a small event and accelerates rapidly until the expulsion of the baby ends it abruptly a few hours later.
Probably at this stage you guys are about to switch off and link to something more interesting on YouTube, but stay with me because I’m about to get to the horse part.
I got a phone call yesterday from a lady who was quite upset because her horse had bolted on her and ran through a fence and into a ditch where he somersaulted onto his back. She and the horse had minor injuries, but she was very shaken and I guess her confidence is badly bruised. She described her horse as normally a quiet and reliable type that responds to the aids very well. The bolt took place over about 100m and she tried everything to make him stop. She tried turning him, pulling with both reins – but he did not listen. The lady felt her horse did not even know the fence or the ditch were there when he hit them.
With a horse (and people too), the reaction to panic has 3 stages. The first is a rapidly rising response to fear. The second is when the fear response is no longer increasing, but is steady. The third stage is a diminishing response to fear. So the reaction to fear in a panic is bell-curved.
The first stage closely mimics a positive feedback loop. This stage can be quite brief or drawn out, but it always involves an increasing reaction. As it continues it usually accelerates until something happens to interrupt it. In the case of the ladies horse, it was going to take something huge to break the positive feedback cycle. The horse did not respond to the rider’s reins. It didn’t respond to the fence as a barrier. It didn’t even respond to seeing the ditch. It took being flipped over on it’s back to break the reaction. This is a true panic or bolt.
Many of us have ridden horses that we think have bolted in a panic, but they often stop when they are far enough away from the cause of the panic. This is not a true panic or bolt. It’s just a serious reaction to something fearful. The difference is that in a true panic or bolt the horse has no conscious control of his reaction. But in a non-panic situation the horse is choosing how to react. In a non-panic situation a horse can be trained to make other choices of behaviour. He doesn’t have that luxury in a true panic.
There would have been nothing the lady could do to stop her horse and keep them both safe. She either had to bail out or ride it out until either exhaustion or a catastrophic event decided the matter.
I have worked with a couple of horses in my time where I was sure they were in a true panic. In fact, one of them seemed prone to it because she would succumb once in awhile. She was scary to ride and in the end I decided she wasn’t safe. She was going to kill herself and a rider one day.
Most horse training works on a exploiting the negative feedback systems of a horse. In the 1950s, BF Skinner wrote about negative reinforcement as a training paradigm for many animals. In reality, Skinner theories were just putting a psychological twist on the well-known physiological phenomena of negative feedback. Both are systems for regulating the function of a biological system. With negative feedback you might regulate the body temperature, but with negative reinforcement you might regulate the stop/go behaviour. They are very similar in principle.
I’m sorry if this topic has been a bit too theoretical and academic. I find it interesting, but I imagine not too many others do. Next time I’ll try to write on a topic a little less theoretical.