Here is the SciBlog discussion:
http://scienceblogs.com/tetrapodzoology ... sonous.php
Oh, and the brain-part I'm thinking of is the neocortex.
Plus, my error, it doesn't say the "forget your dreams" idea is officially discredited, although I thought there was a link to another source that cast doubt on this.
Anyway, echidnas and platypuses (platypodes) dream more than any other mammal, and it's likely this is not because they have to forget a lot of things, but because they have to process a lot of info, which dreaming helps them do.
The platypus apparently dreams more than any other mammal, and engages in lots of REM sleep. Echidnas do this too, despite early reports to the contrary (Siegel et al. 1998).
There sinks the hypothesis that echidnas had such a huge neocortex because they had no REM sleep and were therefore unable to get rid of old informations. I had championed this once, but I have to admit defeat now.
This leaves the Question - why the huge neocortex? Most other ant- and termite eating mammals like anteaters and pangolins are small-brained (I don't know about the aardvark) and seem to get along well enough with it. Breaking into an ant- or termite nest needs big claws and lots of muscle, but it isn't much of an intellectual challenge. Most other australian animals have smaller brains than animals that hold the same niche in other continents - obviously because Australia, a dry continent without much vulcanism, is a low-energy environment. So why is it the other way around with anteaters, pangolins and echidnas?
This leaves the Question - why the huge neocortex?
That's what's called a "good question". We don't even know why we have such a big one. There's a girl in Germany whose, I think, right hemisphere is missing. It was taken out due to cancer or something. Tomography images are published: one side of the skull is empty, filled just by cerebrospinal fluid. Now, apart from a few oddities, said girl is normal, plus bilingual in German and Turkish. We have no idea what a bloated brain is good for.
BTW, ours isn't that big either. It's at the upper end of what's expected for a mammal of our size -- not beyond.
In any case, the speculation that the purpose of dreaming is forgetting was rather silly anyway. "The closer you get to humans, the worse the science gets."
Well, I thought the Crick and Mitchison dream theory was pretty nifty [Warren, I think this may be the theory you are talking about], and is not done justice by such tags as 'reverse learning' or 'the purpose of dreaming is forgetting'. If you read up on neural network modelling of pattern recognition you get an idea of just how much training on positive and negative examples it takes to get a reasonable level of accuracy with new stimuli; also the 'parasitic modes' C&M talked about are a real issue for models, and also make intuitive sense as a source of tics, compulsions and obsessions as well as bad dreams. You may be right, David, but so far all the science we have is done by humans.
> Is it possible that the large brains of modern monotremes is a
> relic of a common ancestor?
If this common ancestor was indeed aquatic or semi-aquatic, this would make a lot of sence; living in the three-dimensional environment of a water collumn and having a sence of electroception would probably be a catalyst for large brain size (mormyrids are an example).
Apparently echidnas are remarkably good at finding their way through mazes and similar tests, so not only do they have big brains, they use them too.
Anyway, I've been thinking about the echidna's large neo cortex and it occured to me. Might the large neo cortex allow an echidna to construct and maintain a mental map of his territory. Much as the large neo cortexes of most primates allow them to construct and maintain mental maps of their territory.
That is, an echidna is able to remember where things are in its territory, and to modify that map when things change. Not how fascinated by that camera the echidna in the one picture is. It is thoroughly investigating the object so it can recognize it again should it come across it later.
Dreams = subconscious learning = pattern recognition in 3 dimensions.
A recent psych. study discussed that: Questions could be solved in either a hard method or an easy method, those that took a nap solved them using the easy method, those that stayed awake couldn't figure out the easy method, although they got the right answer. Dreams are more efficient (time-wise) learning.
One reason I don't like the "dreaming is for forgetting" hypothesis is what I dream. Whenever I can remember it, it's vaguely based on something that I'm not necessarily supposed to forget (such as rising earlier the next day) and builds it into a completely wacky story.
Another is that the opposite of forgetting sometimes happens in my dreams: reinforcement. I've spoken foreign languages and done simple calculations in a few. I also often try to read, especially scientific articles, but it's very frustrating because the text keeps changing in front of my eyes.
Furthermore, I do forget stuff during the day. This especially concerns having done automatted procedures like locking the door and taking the key with me: I can't remember having done it, yet I have done it.
And then there are people with a photographic memory who seem to forget nothing at all and yet have space for all that information in their brains...
In sum, I like the hypothesis best that says dreams are "thunderstorms in the brain", association and imagination running free and unconstrained. If they have a positive effect, it's the joy of thinking, though my dreams are almost never pleasant (usually they aren't nightmares either, though, just boring and illogical).
"I don't know if you can call it a stereotype when I was in a room full of people actually doing it." -- Keith S.