Friday, October 17 2003

Jason mentions

how his brain went on autopilot while he was starting a weblog entry.

This reminds of something that was happening to me while in the UK. The Xerox offices in Bourne End we’re very open – no cube walls, all desks connected. I really enjoyed the social aspect of it, and ended up getting more done when I was surrounded by people. It had one interesting side effect though:

If I was deep into writing code and someone came up to talk to me, while turning my attention to them, and listening to what they had to say, I’d often try to complete the “thought” I was writing. Often what came off the tips of my fingers was fascinating. It would usually be code, but the code would usually be nearly completely random snippets of bigger pieces that I’d recently written, or just total jumbles of code that had no place being there.

Over time I’ve noticed that I often am able to write complex code without exactly thinking about syntanx or anything. It just flows out as easy as chatting with someone or writing up an email, or even walking.

After watching a recent Scientific American show with Alan Alda, I’m convinced that when we continously repeat similar tasks, that the part of our brain that processes that task moves from the more sentient, “human” part of our brain to the more primitive, auto-pilot part of our brain. The frontal lobes handle new experiences, and new problem solving, and essentially train some of the rear lobes to do the task.

Tests on toddlers on this same show illustrated this very clearly. Toddlers were tasked to sort cards with red stars and cards with blue stars into one pile, and cards with red trucks and cards with blue trucks into another pile. When the rules were changed, and the toddler was asked to sort by color only (all red cards in one pile, etc), they failed over and over again. They had a very difficult time untraining themselves from the first task.

Toddlers’ frontal lobes are far less developed than adults. As I watch my daughter learn and grow,  its quite obvious that her problem solving skills are just now merely emerging. I can get her to make the same mistakes when she’s playing with her shape sorting toy as the toddlers in the show. She can put five circles into the toy, but the second you give her a square, she goes straight to the circle hole (even though she’s been able to get the square shape into the square hole before). Show her how to put the square into the square hole, and she will keep going for the circle hole. However, given enough successes with the square shape into the square hole, and she’ll become an expert again, failing the second you give her a circle. As her brain develops, she’ll begin to further understand the role that the individual shapes play in this exercise, but for now, her instruction set is simple: this object goes in this hole, repeat.

It makes sense that our brain behaves like this. It’s like our brains are executing common functions in a less “CPU” intensive area of our brain (the rear lobes). A part of our brain optimized for repetition. The code is written in the less primative part of the brain (frontal lobes) optimized for problem solving, the part of the brain that memorizes the instruction set and prepares it for the repetitive part of our brain.

If our brains didn’t do this, we’d basically be monkeys. If we only had the repetitive part of our brain, we’d have little room for problem solving.  If we didn’t have the repetitive part of our brain, we’d be overgrown newborn babies, having to address every problem as a new problem. Granted, there are other parts of our brain that handle basic motor skills, body functions, etc, but the parts of us that are essentially human would be overworked.

On a energy conservation level, it makes sense to spend some effort solving something, then hand it off to a less intense part of the brain. Our bodies are machines built to be as effective at conserving energy as possible. Using just the problem solving part of our brain to function would be like trying to use the first gear in a transmission to get up to top speed. It might be possible, but in the process it would burn a lot of energy, and put the entire system at risk for failure because of the stress.

The human brain is simply amazing. Yet, with billions of human brains roaming the planet, we still haven’t figured out how to make a good pair of nanopants, with built in, gravity and motion aware Insta Seat Deployment. That sucks!