Locked in, unable to move or speak

16 05 2016

The two weeks of Presentation2of May were packed with local meetings (I missed the OEEEC this time: just too much going on), including SONA which this year was at the University of Waterloo.

As for the Logic meeting, I pared my attendance down to just one plenary, but was so glad I made that. Adrian Owen (below) presented on his work with comatose patients (people in vegetative states and comas at one end of the range — where the lights really are off  — and locked-in syndrome at the other: fully conscious but unable to do a thing). It was one of the best talks I’ve ever seen in my life: it actually made me physically happy to be listening.

He opened by talking about some of their work trying to predict which of these patients will recover. Using fMRI, they look at which brain regions respond to what type of auditory probe: non-linguistic sounds, linguistic sounds that are nonsense, and linguistic sounds with meaning. All of these activate the auditory cortex, and that doesn’t predict a thing: those cellular responses just reflect that the sound has made its way in, but no more – they don’t require or anticipate consciousness. Linguistic sounds activate other cortical areas too (e.g. Wernicke’s area) in some subjects, and those in whom this occurs have a better prognosis. What I missed was whether responding to real speech has better predictive powers than jabberwocky, but I think so (need to read Davis et al. ’07 in PNAS, and Coleman et al. ’09 in Brain).

Since being able to follow an arbitrary commend (“Lift up your left hand and wiggle your fingers”, that type of thing) is use by clinicians to assess awareness, another body of their work used this approach in their patients. Of course these people cannot move, but some of them can imagine moving when asked to. And just as our hippocampus would ‘light up’ if we imagined moving around our house, so too can theirs; and just as our premotor cortex would ‘light up’ if we imagined making a tennis swing, so too can theirs. Patients who can do this have a good prognosis (and are fully aware – later, once recovered, they can describe how they felt taking part in this experiment). Interestingly though, as with all their techniques (he admitted) even this is prone to false negatives: some patients who later recover and can recount the whole experience, simply cannot pass this imagining task.

With the subset of patients who can do this, things now can get really clever. They can be asked to imagine walking round their house when the answer to a question is ‘no’, and playing tennis when the answer is ‘yes’. A few test questions are used to check they’ve got it and it’s working (“Does Margaret Thatcher run Canada?” “Are bananas yellow?” – that kind of thing). Once that’s been confirmed, then they can be asked questions the questioner does not know the answer to, such as “Are you in pain?”, and “We’ve been playing you your favourite album for seven months now: do you still like it?”. (Owen et al. ’06 in Science and Naci et al. ’14 PNAS look to be good follow-up reading)

Incredible, incredible stuff. Oh and Adrian Owen is a Brit, so of course I lerrrved his accent.