by Sven Triloqvist
Mon Nov 26th, 2007 at 03:34:58 PM EST
There I was reading a BBC business item about the role of mathematicians in market mayhem, and learning how much these quantitative analyst types make from their little models, when blow me down, there was this...
When I telephoned one fund to ask if we could interview some of their quants, the receptionist told me that I was unlikely to get permission, partly because of the secretive, commercially sensitive nature of their work but also, because many of them in that fund are autistic.
A strange destinationless journey below the fold...
Now a good friend of mine has an autistic child, so I am well aware of both the tragedy and the consequences for the family. But for those who may not, here is the wiki link.
Autism is a brain development disorder that impairs social interaction and communication, and causes restricted and repetitive behavior, all starting before a child is three years old. This set of signs distinguishes autism from milder autism spectrum disorders (ASD) such as Asperger syndrome.
Now a rare occurrence in autistic people is savantism - having a very narrow, but remarkable intelligence that can manifest itself, for instance, as an amazing ability to do arithmetic calculations. This however is not so in the case of my friend's child.
An autistic savant (historically described as idiot savant) is a person with both autism and savant syndrome. Savant syndrome describes a person having a severe developmental or mental handicap with extraordinary mental abilities not found in most people. This means a lower than average general intelligence (IQ) but very high narrow intelligence in one or more fields. Savant syndrome skills involve striking feats of memory and arithmetic calculation and sometimes include unusual abilities in art or music. Savant syndrome is sometimes abbreviated as "savantism", and individuals with the syndrome are often nicknamed savants. This can be a source of confusion since a savanter is a person of learning, especially one of great knowledge in a particular subject.
The origins of autism are not fully understood - but there are some clues, and genetics plays a major part. There appears to be a connection between OCD (Obsessive Compulsion Disorder), pathological gambling and autism. Normally disorders are divided into 3 neurobiological categories: Dysmorphism (a preoccupation with body as seen eg in anorexia), neurological disorders (autism, Tourette syndrome) and repetitive behaviours (driven by pleasure or arousal, such as sexual obsessions, kleptomania, pathological gambling).
A couple of day ago a new piece of the jigsaw was added: Brain scans 'may detect OCD risk'
These brain changes appear to run in families and may represent a genetic risk factor.
This was associated with decreases of grey matter in brain regions important in suppressing responses and habits - the orbitofrontal and right inferior frontal regions.
Researcher Lara Menzies said: "Impaired brain function in the areas of the brain associated with stopping motor responses may contribute to the compulsive and repetitive behaviours that are characteristic of OCD.
Coincidentally, increased metabolic activity in the orbitofrontal regions appears to be associated with anorexia. And families producing autism also have a high instance of OCD ie the genetic component.
Medical trials have also identified the inhibition of serotonin uptake as a possible treatment. A genetic inability to produce sufficient quantities of GABA will also affect suppression of repetitive behaviour. GABA is the biochemical that drives the inhibitory neurons that clamp down on neurons that are firing repeatedly. (GABA runs out during the day and has to be replenished at night, as we sleep).
I have also commented on the opioidergic system before, though for the life of me I cannot search them out at the moment. Sorry. In brief, the release of hormones, semi-hormones and neurotransmitters into the interstitial brain metaprogram the functioning of the neural networks. The effect of these releases is usually temporary (ie noradrenalin in response to perceived danger), but many of these `metaprograms' have long term consequences: for instance, endorphins that are released as a consequence of alcohol drinking, enter a receptor and promote in the firing neurons that they attach to the creation of more connections between other neighbouring neurons that are firing at the same time. That is how alcoholism gets hard-wired.
Now we get into more speculative territory...
Two things seem to be going on:
1. Genetics will decide the relative capabilities of the `manufacturing centres' in the brain. That is, how much (if at all) of these hormones, semi-hormones and neurotransmitters will be produced, and then removed by uptake. These might be described as the rules for a self-organizing system. You knew I`d get to that at some point ;-)
But 2. The `triggers' for the release of these metaprogramming biochemicals can be inherent or LEARNED in the process of self-organization of neural networks. The inherent ones are those essentially more basic triggers for eating and sleeping, for instance, that involve internal feedback from the presence of `gradients' of chemicals. Hunger is a gradient. The longer the trigger is present, the less likely it can be consciously suppressed. However, in the case of anorexia, there appears to be a recursive suppression of the trigger: if your metabolism is skewed to being thinner, not a finite quality of `thin', then it will be recursive.
This recursiveness seems to be a feature of brains out of balance. The more you drink alcohol, the more you are likely to drink alcohol in the future - etc.
Recursiveness can be seen in the apparent connection between low self-esteem and overeating disorders. The more you eat, the lower your self-esteem, the more you eat.
These are Learned Behaviour Disorders. They are a function of the metaprogramming. Sometimes metaprogramming totally fucks up. Here is the case of a swan in love with a paddle boat. Of course we'd never be so stupid ;-) (says the man who has been divorced twice)
Somehow I still have to get back to the autistic quants. Bear with me.
Firstly though to reiterate: in a self-organizing system like the brain, there are `rules' that govern the biochemical microcosm at the cellular level and the molecular level. The `actions' of the brain (stimulus/response) are `controlled' by these rules from the very beginning of life and its confrontation with the massive undifferentiated noise of stimuli.
The lifelong process of differentiation - getting signals out of the noise - is similarly controlled by these rules. BUT the experience of life can change the trigger mechanisms for the metaprogramming. Or can it? Is the thrill seeker addicted to noradrenalin because he/she produces more of it naturally in response to the trigger?
And so to quants. A hedge fund, or indeed the whole industry of financial legerdemain, is a system that is now bionic. It quantifies risk by the output of mathematical models run on computer systems. The computing system itself can trigger trading activity, or rather the model can. In fact whether the system does the trading or whether a human does the trading is partly irrelevant: it is driven by the results that the model spits out. What do the humans do? Well they obviously spend a lot of time creating the rules that `control' the system to ensure that it interfaces with the law. And they also have to sell their skills (and their model) to investors. But are they involved in trading decisions, or just in the construction of (il)legal packaging? Some professional will no doubt fill us in here.
To me, though, the interesting thing is that the moral aspects of trading seem to have been subsumed into the model-makers and their models. Neither of which, if we are to believe the existence of these autistic quantifiers and their modeling, are fully connected to ideal social behaviour. Don't get me wrong - I am happy that these savants are employed, but their disconnects from social behaviour are well documented. I just think their unique and narrow skills should be better employed, especially when their current employers show a variety of OCD disorders - including pathological gambling.