What good science looks like

Yesterday, I turned in a rather lengthy paper for chemistry class, the end product of I forget how many hours going to libraries, accessing online journal archives, figuring out what specialist-oriented articles were saying, and then typing the thing up. It's something I think everyone should do, and not just because I'm a sadist who believes that if I must suffer, everyone else should too.

The assignment was to find an article in the popular press dealing with science and reasearch the science behind it by going through peer-reviewed journals. My article was from Scientific American, on memory enhancement in humans. I focused on the reaserch on NMDA receptors by Joseph Tsein and Ya-Ping Tang:

Before a neuron naturally increases CREB, certain channels on its membrane must open to allow positive ions to flow into the cell. The ions then trigger a cascade of events leading to the activation of CREB. One channel of interest is known as NMDA. In 1999 Joseph Z. Tsein, Ya-Ping Tang and their colleagues, then at Princeton University, discovered that increasing the number of NMDA receptors in the mouse hippocampus led to better performance on a spatial-memory task. Now researchers and pharmaceutical companies are pursuing NMDA receptor agonists (they combine with the receptors) as nootropes. At least a dozen new drugs of this kind are making their way toward clinical trials.

The pharse "One channel of interest" hints at a larger field of research, but one might be forgiven for thinking these were some really smart guys who made these super mice all on their own.

The truth becomes clear with a glance at the "references" section of the paper, in the September 2nd, 1999 issue of Nature where they published their results. It's 30 items long, some of it involving the same scientists who enhanced the mice, but mostly other people's work. What's more, some of the things they cited were review articles, scholarly articles that bring together the work of many people in one place.

Furthermore, some of this stuff goes back a long way. The general principle they worked with, Hebb's rule, was proposed in 1949. The more specific process, called long-term potentiation, was first described in 1973. The earliest review article talking about NMDA receptors I found dated 1981.

In the main paper, all the researchers had done was take an existing technique and used it to enhance an already identified gene, then test the resulting mice. This was, in other words, not a sudden brilliant accomplishment, but a small finishing brush stroke on a large body of research. Perhaps finishing is the wrong word, though, as this research will be built on further as companies try to develop drugs based on it.

This reminded me of one of Bob Park's warning signs of pseudoscience:

6) An important discovery is made in isolation. Most scientific advances draw heavily on research by a number of scientists or groups working in related areas. Successful innovators tend to be actively involved in the open exchange of scientific ideas and results, presenting their work at scientific conferences and publishing in mainstream scholarly journals. The image of a lone genius working in secrecy in an attic laboratory who makes a revolutionary breakthrough, is a staple of Hollywood horror films, but its hard to find examples in real life. There are frequent claims by lone inventors to have made such breakthroughs, but the claims rarely if ever stand up.

Good science then, builds on other things, and is built upon. In the Nature coverage of Tsein's research, a key point was that it helped confirm certain ideas about the workings of the brain. When countless pieces of careful research have built on an idea, each helps confirm it, and though no piece may provide any degree of certainty on its own, taken together they can constitute a nearly indubitable proof. Atomic theory and the theory of common descent are like that. Intelligent Design is the polar opposite; not a shred of research has been done based off of it. Parapsychology is better, but only to a point: they keep doing initial experiments, but it never builds.

Unfortunately, it will take quite a bit of work to get people to realize this. As I mentioned in my introduction, one might think, from reading the Scientific American article that this was the work of lone geniuses. Similarly, many people probably think that the evidence for evolution began and ended with Darwin. We might, though, work on this problem by inviting people to browse the scientific papers behind a random news report. We might, if not for the strange looks doing so would invite.