Lessons from Northstar's botched study of a brain implant

northstarbrain_white.jpg I spoke with John Bowers, the chief executive of Northstar Neuroscience, about the electric brain stimulation trial Northstar will launch this year for the treatment of depression. This small medical device company hopes to soon market its brain implant, which sends pulses of electricity from a device inserted in the neck to an electrode that sits on the outside of the brain, just underneath the skull.

Up to one-third of patients fail to respond to conventional antidepressant drugs, and a number of medical device companies are exploring the use of brain stimulators--both implanted and external--to help correct problematic electrical activity inside mood-controlling regions of the brain. But Northstar started off 2008 on the wrong foot: a study of the same device used in stroke rehabilitation showed it wasn''t any better than intensive physical therapy.

Ever a fan of neural prostheses, I asked Bowers what we could learn from the stroke trial to improve the one for depression. Here's what he had to say.

1. Include more patients. ''You''ll probably see us do a much larger number of implants than we did with stroke,'' Bowers says. It''s not uncommon for a number of severely depressed patients to equally fail to respond to antidepressant drugs but to each exhibit different sets of symptoms and behaviors, meaning that depression doesn''t follow along one simple pathway in the brain.

2. Pay extra attention in choosing the comparison group. The durations of these clinical trials are usually several months, often a year--far too long for patients to go without any treatment. So what''s the best therapy for the group that doesn''t receive electric stimulation? The choice is always difficult.

3. Make sure the chosen patients are really, truly resistant to antidepressants before getting into the heart of the study, by building in an 8-week pre-trial period, during which patients will receive other treatment. If they improve, the subjects should be excluded. Once the subject pool has been refined, the chances of observing a placebo effect are much slimmer. ''These patients will have already failed 9 or 10 therapies, so their hope in a new one is already low,'' Bowers reasons.

4. Don''t take the stroke trial too seriously: ''We''ve always said that the stroke trial couldn''t be a predictor for the others, even before it started,'' he says. ''The basic hardware is similar, but the treatment algorithm and stimulation parameters are different.''

I have my own reservations about the upcoming study, but I''m not a medical device company, so I don''t really know. [[CORRECTION APPENDED]] Northstar believes the outcomes of the depression trial will be positive enough to enable them to apply for FDA approval for the device. But you may wonder how they configured the device, given that they were, presumably, equally confident in the stroke trial. The settings of the depression treatment algorithm''stimulation frequency, duration, target location, and so on''were refined in the lab and tested on 11 patients in a preliminary human study.

But is data from 11 people really enough to work out all the kinks? To be sure, a solid body of research supports the company's choice of the brain region to stimulate, the dorsolateral prefrontal cortex. But to me, basing a company''s''and a therapy''s''success on results from 11 people seems like quite a gamble.

CORRECTION: A Northstar rep tells me that the depression study will not culminate in FDA approval and is a more preliminary study of the technique's feasibility.


Tech Talk

IEEE Spectrum’s general technology blog, featuring news, analysis, and opinions about engineering, consumer electronics, and technology and society, from the editorial staff and freelance contributors.

Newsletter Sign Up

Sign up for the Tech Alert newsletter and receive ground-breaking technology and science news from IEEE Spectrum every Thursday.