notes-science-neuro-managementSecretsOfTheBrain

comments sent to an email list on the article "management secrets of the brain"

http://www.business2.com/articles/mag/0,1640,43499,00.html

interesting paradigm and one that I think will be very useful in business once us neuro guys start to figure out more of what is actually going on. It is true that multihuman "agents" like corporation share many of the same problems as the brain (information processing, decision making) (although they are definitely fundamentally different; for instance, the brain doesn't have to worry about motivating agents with truly competing interests (at least on the level of small sets of neurons, whether we have actually competeing subpersonalities is open to discussion in my view)).

also, i think this kind of thinking is interesting for us neuro folks too by giving us analogies with which to restructure our thinking.

> 1. Never try to micromanage a large, complex organization.

although one of the things i think the brain can do that human organizations can't do so well is selectively micromanage -- not down to the level of perceptual processes, but down to the level of "cognition". if businesses were more like this then there would be very few instances of bureaucracy that is obviously ineffecient from the point of view of the affected workers but which wouldn't go away.

> 2. Don't let bottom-up self-organization go wild.

I'm not convinced this is a good prescription for business (at least not in the way that the author meant it; i.e. hierarchy is needed). and as for the brain, all we can really say at this point from examples like Gage is that there are modes in which mental subprocesses, when left uninhibited, are detrimental. we haven't yet established if the control structure is entirely bottom-up or not.

in my view the role of hierarchy in business is mostly due to humans being used to hierarchial social forms. The only way I can see that it might be NEEDED is for political reasons (ie without a real hierarchy people will be jockying for position anyway, possibly in a detrimental way).

> Without leadership, after all, standard operating procedures are > directionless and blind.

this is the statement that i'm disagreeing with above

> But how? Until recently, says Carnegie Mellon University psychology > professor James McClelland?, most scientists believed that the answer > would be very complicated. Their basic idea was that whenever your > automatic processes came into conflict or whenever you needed to behave > in a new way, the prefrontal cortex would instantly take over.

but we can do this; if you read a description of how experts say that you should organize your thought process while playing chess, you are able to do it yourself. this goes somewhat further than just encouraging/discouraging subagents according to some sort of "is this working?" criteria. in other words, you are able to form and implement declarative knowledge about the way your brain procedures work. like the Java programming language "reflection" or "introspection" ideas.

> But this > approach is micromanagement in the extreme -- the neural equivalent of > the front office instantly sending in a Tiger Team to replace the paper > in the photocopier. It's also grossly inefficient. "You'd have to > maintain two completely different sets of abilities in the brain," > McClelland? says.

the actual abilities and the meta-level (reflection/introspection) abilities? perhaps. but perhaps you can use the same representation and information processing for the meta-level abilities as you used for your other abilities. so you wouldn't necessarily need a totally different architecture.

> Just within the past decade, however, some neuroscientists have begun > to advance a simpler explanation of cognitive control. Princeton's > Cohen, who is one of the strongest advocates of an alternative view, > offers a corporate metaphor: The prefrontal cortex simply communicates > its intent and leaves subordinates to work out the details and > execution. This would seem to imply that ... .... > This new model of cognitive control assumes that the prefrontal cortex > has just one job, which is to generate a neural map of the brain's > goals, strategies, and current situation. Call it a mission statement. ... > Got an itch? You scratch. But if it's caused by a mosquito bite, a new > activation pattern is broadcast to suppress your immediate impulse to > do something about that itch. In general, Cohen says, the prefrontal > broadcast will bias your various impulses, strengthening some while > weakening others, so that the desired activity can compete and win.

this is an interesting idea, and one i think is mostly right. i don't think it is in direct conflict with the "old model" that i was defending above, though. hypothesis: we can introspect to an incredibly specific degree, however, most of the time we don't and we use this agent-prodding approach instead.

of course in reality the agent-prodding and the introspection is probably implemented by the same architecture; it seems agent-prodding is more likely to be able to cleanly emulate introspection than the other way around, so maybe there is some sort of superiority to this view, even if both processes happen.

another hypothesis might be that we do agent-prodding for everything, but that we can reason about our agent-prodding, i.e. "i want to think methodically about chess, and to first consider this postition, then that one; so first i'll inhibit this agent, and then that one". I bet this goes on to some degree, but it sounds pretty complicated to me so i doubt this is the feel of the way that things really work (and it looks like there is some recursion problem there; how could you reason so well about how to prod these guys without having a complete, functioning mind in the first place?)

another way to put that is, whatever is doing our executive functions has so much reasoning ability that I doubt that it's only allowed method of interaction with everyone else is a "mission statement".

> 3. The best way to control your subordinates is to just point them in > the right direction.

perhaps true for humans in current social modes, but I think the brain can do better, as i said above

> 4. Be careful listening to the voice of experience -- that voice could > be your own.

I don't quite understand this prescriptive. Is he saying that business leaders should look critically on what they consider to be empirically proven information, because the empirical "proof" could actually rest on their own biased interpreations of their own past experiences?

> That aide would have to keep outside disturbances at bay as long as the > prefrontal cortex's strategy is producing the desired results, but let > outside information come flooding in when things start to go poorly. ... > intellectual reward system. The dopamine system can also disrupt the > prefrontal cortex with inappropriate input, which can explain many of > the symptoms of schizophrenia, including the victim's disorganized > thinking and inability to concentrate. (We wouldn't want to make this a > management rule, but it would appear that when the CEO gets in trouble, > the blame falls on the secretary.)

this would be a fun management system; when the company is doing well, the secretary does what they are told, but when the company is doing badly, they refuse to screen out random blips of information (i.e. they let any schmoe make an appointment with the boss), under the assumption that the CEO must be living in some sort of fantasyland and isn't getting the information that they need to hear.

> Which suggests that ... > > 5. The organization can't succeed without passion.

well now that's a stretch!

-- bayle