January 16, 2012
Mathematical Power Law For Serial Killer Behavior
The Technology Review Arxiv blog has an interest post about how a Russian serial killer's frequency of killing fits a power law that suggests a pattern of neuronal recharge after killings.
On 20 November 1990, Andrei Chikatilo was arrested in Rostov, a Russian state bordering the Ukraine. After nine days in custody, Chikatilo confessed to the murder of 36 girls, boys and women over a 12 year period. He later confessed to a further 20 murders, making him one of the most prolific serial killers in modern history.
Today, Mikhail Simkin and Vwani Roychowdhury at the University of California, Los Angeles, release a mathematical analysis of Chikatilo's pattern of behaviour. They say the behaviour is well characterised by a power law and that this is exactly what would be expected if Chikatilo's behaviour is caused by a certain pattern of neuronal firing in the brain.
Click thru and read it to see if you find their conclusion likely. It seems plausible to me. Here's the abstract.
What I'd like to know: Short of committing a real murder how to drain the neural excitation that initiates a serial killer's desire to kill? On a related note: Has the playing of violent video games reduced the frequency of killing by serial killers? Could we detect the draining effects of simulated kills in video games by looking for signs of lower frequency of serial killing over the last 20 years?
What would really help: better ways to detect serial killers: What sorts of measurements could detect a person's propensity to become a serial killer? I've been reading a number of books lately about psychological research (e.g. Daniel Kahnemann's excellent Thinking Fast And Slow) and am struck by tricky means psychologists have devised for measuring cognitive phenomena. For example, pupil dilation happens and can be usefully measured when the conscious brain is thinking hard. Can any existing tools of psychological and neurobiological research show differences between serial killers and the population at large?
We need the ability to detect serial killers in advance. How to get from the idea that neurons charge up to activate serial killing circuitry to a way to more rapidly detect serial killers in our midst? Ideas?
If there is a refractory period after such an urge is satiated, then it is possible that this theory has some validity. Even animals take time to "recharge" after the buildup and release of a normal, healthy behavioral pattern. There was a pit bull that mauled a 1 year old, in the news, over the past day or so. When the cops arrived, the dog was still in a high state of agitation and was shot dead by the police. Cats that are threatened take a long time to cool down before you can play with them again. Some animals do certain behaviors repetitively, such as reproduction, except they don't do them one after the other in short intervals, but sometimes the periods are separated by months or years. I think there may be some validity to this theory and it would be interesting to see if there is more data to support this hypothesis.
Too bad that Numb3rs is no longer in production. My wife really loved it.
You see where this is heading, don't you?
Government will decide that the only way to "save the children" is to get into our heads and monitor us, to make sure we behave.
Get ready for the IEPA (Internal Environmental Protection Agency).
You think I'm joking?
serial killers are natures way of removing the dross out of society. most victims of sk's would never have made a significnat contribution to the betterment of mankind anyhow. in the human body killer t-cells kill foreign antibodies. seria killers are the human races version. Sorry if thats too controversial a thing to say.
Power laws show up everywhere. While it might be suggestive, it's only that. As an example, my consumption of any given food that I like follows a similar power law trend. I eat it, then the sensory experience persists afresh in my memory for some time, making it less desirable to repeat that meal. Instead, during this "refractory period" I pursue some other foods until I'm again in the mood for the one I ate previously. Would anyone seriously suggest that there's a specific neural circuit for each food, for instance, a chicken-fried steak circuit? No, this behavior operates on a conscious level according to preference for novelty or variety (which may have an innate component).
Furthermore, it could be that serial killers would prefer to kill non-stop, but they are discouraged by the increased "heat" and fresh leads generated by law enforcement in the wake of a new killing. The serial killer merely waits for the scrutiny to die down or for the leads to grow stale.
As for predictive intervention against possible serial killers, I just don't see how that's compatible with a free society. I think a better approach would be accountable, automated surveillance whereby a system with documented algorithms that can be freely audited becomes solely responsible for real-time processing of gathered surveillance data to ensure public safety. By taking away human discretion and secrecy in the review of surveillance data, and by making the system's inferences and methods public, much of the potential for abuse of individual rights is preempted. This sort of expert system could simultaneously monitor far more criminal threats than manpower would otherwise allow, and it would be incapable by design of extending the scope of its inquiry beyond imminent threats to public safety. It would be a kind of benign and non-judgmental watcher that only looks out for the protection of citizens and disregards the rest of its input.
The model is wrong (http://cscs.umich.edu/~crshalizi/weblog/857.html), the data don't fit a power law.
From the arXiv pre-print: "Thus, one may speculate that similar processes in the brain may lead to both epileptic seizures and serial killings"
That says it all, the key word is 'speculate'
As an aside, there is a cottage industry of power-law 'findings' in many disparate fields. Many (if not most) seem to be bogus -- Shalizi has written extensively about this.