Author: jlolds

Native Angeleno, professor of Neuroscience and Public Policy at George Mason University

A dark view of America’s future…

Here in FT’s magazine, by Edward Luce. At one level, it’s overly simplistic because Luce is trying to make a point–namely that the US is in decline across multiple dimensions. The truth is, and to be fair Luce acknowledges this point, this country remains a leader in many areas. But the piece makes many excellent points, particularly with regards to the political paralysis that I’ve blogged about recently.

But America has been at such a point before (think 1861 or 1968). Historically, it has shown an ability to renew itself at many decisive points and frankly the society is too complex to forecast any particular outcome with any degree of certainty.

In my own travels I have been impressed with how much innovation and creativity remains here in America. And further, the increasing productivity of humans may not be zero sum.

Hope for Congress?

About 10 miles away, over the rainbow, is the Capitol Building. Perhaps they will overcome their differences in the near future and relearn the art of compromise?

In all seriousness, political failure is now a real part of the calculus for assessing systemic risk here in the US. My own sense is that, after the general election, enough progress will be made on the country’s economic trajectory to keep the recovery going–but nothing is certain and the two sides are playing very much of a zero-sum game.

Personalized medicine "omics" approaches hit a speedbump

ScienceInsider has the story here. Money quote:

One major problem, the report says, is “overfitting”: Because the studies often look for patterns in hundreds of biomolecules using a relatively small number of patient samples, it is easy to find correlations that do not reflect the biology of patients’ disease. The report recommends a set of steps to validate the tests, such as repeating the test on blinded samples from a different institution. Journals and funders should also require that data and models from papers be made freely available so that other researchers can check the results.

On the danger of being oversold: NeuroX

I’ll simply note this morning that today’s NYT op-ed piece on interpersonal neuroscience is the latest case of the NeuroX fad–applying a neuroscience framework to any and all social issues. This sort of journalistic conceit is becoming noxious I think–neuroscience as a field is still young, lacks a full theoretical underpinning and is simply not ready to be the explainer-in-chief for all human/animal social phenomena.

The last time a nascent discipline got over-exposed like this was I think the overselling of Artificial Intelligence. It wasn’t good at all for that field.

Postdoctoral education….

Like many life scientists, I see the postdoctoral years as a continuation of scientific training after receiving the PhD degree. Ideally a postdoc should last no more than three years, provide the trainee with the opportunity to acquire several new methodologies above and beyond those learned during the dissertation,  result in at least two new first authorship papers and finally provide grounding in grantsmanship.

A postdoc is most definitely not a super-technician (i.e. a technician with a doctorate). As with the graduate student, there is an important mentor-trainee relationship between the laboratory PI (the mentor) and the postdoc. For this mentorship relationship to work, the PI needs to prioritize the training aspect of the fellowship at the same level as the research component. Often that doesn’t happen and it’s a shame. It should.

Additionally, there is a tendency among many post-docs to view themselves as “research assistant professors”. It doesn’t help that some institutions in fact classify postdocs as such. The problem with this self-image for the trainee is that the education component of the fellowship also becomes de-emphasized (any many PI’s have no problem with it). There may be the further problem of the trainee and PI getting at cross-purposes over who “owns” the research produced during the postdoc.

These days, a scientist may have several postdocs before going on on the job market. That’s OK within limits. Each postdoctoral fellowship should add new techniques, new scientific perspectives and increasing independence. And they should be in different laboratories lest the trainee eventually turn into a super-tech.

FT’s John Gapper interviews MIT’s Esther Duflo

The weekend Financial Times is a real joy to read each week, this one was no exception. John Gapper’s luncheon interview with MIT economist Esther Duflo at a place off Harvard Square was perfect and it’s here.

What really intrigues me is Duflo’s field research methodology. Her conclusions are often counter-intuitive but quite brilliant I think. I’d like to see what she does married up with some of the advanced GIS agent-based modeling stuff that our own folks are doing, see here.