Re: Doogie Mice

Robert J. Bradbury (
Fri, 3 Sep 1999 09:07:26 -0700 (PDT)

On Thu, 2 Sep 1999, Spike Jones wrote:

> > At 12:35 PM 9/2/1999 -0400, ??? wrote:
> > Joe Z Tsie of Princeton University has used genetic engineering to
> > create a mouse with greater *intelligence* than any found on Earth...
> >
> > Philip Witham wrote:
> > How could it be that nature didn't select for this? So simple, so
> > apparently effective, there must be a down side, like - the *brilliant*
> > mouse decides that having children would pinch her lifestyle.
> Exactly! Evidently most ecological niches offer little or no reproductive
> advantage to higher *intelligence*, so the species in those niches get
> no smarter.

I am cleaning out my mailbox, so I wanted to go back and annotate this post. What the original article was about was about affecting long term potentiation and *MEMORY*. We should be careful not to "smudge" memes and excessively extroplate because it ends up with a lot of discussion about something which may be unrelated. Now, I will admit that the article, especially in the abstract implied, that the work could lead to improvements in intelligence, so we aren't completely the source of the problem (though we should do our best to function as accurate reviewers and correct meme-stretching when we see it). I'm very surprised that Science let the article go to press the way that they did (but when it ends up in the NY Times the next day one has a hint as to why).

While a large memory is a factor in intelligence, there are many other factors as well -- perhaps the ability to simultaneously hold multiple concepts (and be aware of them), perhaps the ability to translate in interesting way between audio/video/speech experiences, perhaps the ability to effectively copy an idea from one side of the brain to the other, perhaps subtle differences in axonal architecture that alters the propagation delay of the electrical signal, etc.

Intelligence is this catch-all word that we use to describe a huge collection of processes. I think we are up to 8-12 different types of "intelligence" now (someone who knows this better may want to comment on what they are putatively thought to be), and the LTP/Long-term-memory is likely to be only part of the story in some of them.

Which I suppose is good, since that means we have lots of room for improvement. Generally speaking though, I'd agree with Eugene that the delays in approval processes for genetic intelligence augmentation make it unlikely that we will see much of this before we can put nanobots in our skulls. A key prerequiste will be the gene-chips that allow inexpensive discovery of all of the gene variants (alleles) that contribute in one way or another to various aspects of intelligence. As the brain has more genetic complexity than most other organs -- I've seen some #'s as high as 20-25% of the genes in the genome being expressed in the brain -- decoding it all will be time consuming.