“A new technology could let your computer recommend new music you might like based on an acoustic analysis of the tunes it already knows you enjoy
. By analysing the characteristics of a song – like timbre, rhythm, tempo and chord changes – then comparing it to a database of a million songs, the software can recommend similar pieces of music, and even rank them by characteristics, like their key or dance-ability.” (New Scientist)
Somehow, I think this would be less satisfying and productive than recommendations already available, culled from a much larger database by a far more sophisticated and subtle analytical process! For instance, communities like Audioscrobbler, to which my listening history is automatically uploaded by a plug-in in my mp3 client, will show me music I haven’t yet heard that listeners with similar taste listen to. (I love it that by dragging a slider I can control how obscure or popular the recommendations will be, too.)
Several of the artists on the recommendation list I know to be on the mark, in that I have heard of them and gotten the sense they are up my musical alley, although I have not yet had a chance to listen to them. Several others are names I had yet to discover, precisely the purpose of the recommendation system. I am open to your assessments of them (am I going to like Neutral Milk Hotel? Built to Spill? Destroyer, which sounds like the name a heavy metal band would choose for themselves?) or any other recommendations you might have, based on your appraisal of whom I listen to, by the way…
The only problem I find with Audioscrobbler is that I download alot of music from mp3blogs to try it out, which thus will appear to Audioscrobbler as part of my listening habits although not necessarily stuff that I end up liking. To counteract that, I sometimes keep iTunes playing my playlist of highest-rated favorites even when I am away from the computer to exert a corrective influence on my Audioscrobbler statistics. Weird, huh?
One of the vexing issues in understanding and treating depression is that, although antidepressant medications change the levels of neurotransmitters implicated in depression almost immediately, they do not have clinical effects for several weeks or even several months. Somehow, a change in serotonin levels, say, has to be accompanied by a change in the way the brain responds to the increased serotonin. Now a group led by Paul Greengard at Rockefeller University has found a protein that seems to regulate neurons’ response to serotonin.
In a mouse model for depression which has proven reliable in the past at probing various neurochemical aspects of human depression (the “learned helplessness model”), the protein p11 upregulates the numbers of serotonin-1B cell surface receptors so the cells are more sensitive to available serotonin. The evidence for its pivotal role includes demonstrating that p11 increased in mice in parallel to their response to varied treatments for depression; that mice bred to be genetically p11-deficient are more depressed, have less serotonin activity, and show less response to antidepressant medication; and that mice bred to have high levels of p11 show extra levels of serotonin receptors and do not exhibit depression-like behavior.
The mileage in improving understanding and possibly treatment of psychiatric disorders is all going to come from turning the focus from the neurotransmitter-and-receptor based understanding we have had for the past half-century to an understanding of the involved intracellular processes. I am not sure p11 is ‘the’ answer, since the more we look the more reductionistic we find any given model to be.
However, as I said above, p11 seems to help answer the vexing issue of finding a neurochemical process that mirrors the time course of clinical response to depression treatment. The next generation of psychopharmacology might involve therapeutic drugs that manipulate p11 directly — rather than indirectly through alterations in neurotransmitter levels — to treat depression more efficiently. If a genetic deficiency in p11 turns out to be one of the vectors for hereditary vulnerability to depression, gene therapy to augment the brain’s supplies of p11 could be a preventive measure. I would also of course want to know what else, if anything, p11 does in brain cells, to understand what we could be meddling with in tryng to manipulate it directly.
It is also worth noting that Greengard, whose work I have followed since I knew his son in medical school, shared the Nobel Prize in medicine in 2000 for work which presaged this finding. It seems pretty unusual to award a Nobel Prize so contemporaneously that the Nobelist still has the potential for monumental scientific discovery ahead of him/her.