It is Friday night in New York City. Nearly a thousand people are packed into a theatre, eyes glued to the stage, the atmosphere electric. The crowd isn’t gathered here for a stand-up comedian or musician – instead they are waiting for an answer.
That is because 25 years ago, on 20 June 1998, a neuroscientist and a philosopher made a bet. , now at the Allen Institute for Brain Science in Seattle, wagered that scientists would have discovered the neural basis of consciousness – the set of brain cells that give rise to our subjective experience of the world – by the year 2023. , now at New York University, was dubious. The pair agreed that they would meet again in 25 years’ time, at which point the loser would gift the winner a case of fine wine.
“When you’re young you gotta believe that things are simple,” Koch tells the crowd. It is a recognition that his younger self might have been a bit naïve – and an acknowledgement that he has lost the bet.
Advertisement
Over the past few decades, several explanations of consciousness have sparked excitement but then failed under scrutiny. There are now two main proposals in contention, and the wager between Koch and Chalmers inspired a project called that pits them against each other. Initial results from that “adversarial collaboration” are being presented here tonight at a conference of the Association of Scientific Studies of Consciousness at New York University.
The first is integrated information theory, or IIT, which posits that the consciousness of any system can be quantified. This measure, called phi, is higher in parts of the brain that are more densely interconnected, such as the sensory processing regions at the back of the brain.
The second is global neuronal workspace theory, or GNWT, which zeroes in on the frontal regions of the brain responsible for thinking, attention and problem-solving. Proponents think signals in these regions compete for our attention, and when one signal wins out it is broadcast to the rest of the brain.
The COGITATE project tests key predictions of each model. In one experiment, human volunteers looked at images of faces, objects, letters and made-up characters while undergoing different kinds of brain scans.
The first test was to see where in their brains the features of the images – like the category and orientation – were processed. Here, IIT came out on top: both category and orientation are decoded in posterior regions of the brain.
Next, was the duration of the activity. IIT predicts that conscious perception comes from sustained brain activity, while GNWT says there should be two bursts of activity, one when a stimulus first appears and another when it disappears. Here again, IIT won out.
The final test looked at the patterns of synchronous activity predicted by each model, with GNWT slightly edging out IIT.
In the end, it isn’t a decisive victory for either model, though that hasn’t deterred their proponents. But one thing is clear: we don’t yet have a conclusive neural basis for consciousness. Koch has lost. Conceding defeat, he walks offstage and returns with a wooden box containing six bottles of wine.
“However, I’m gonna double down, and I will challenge you, 25 years from now… we will have clarity,” says Koch.
Chalmers quickly does the maths. “2023 plus 25 equals 2048. You will be 92 years old,” he says.
“A springly 92,” says Koch. “Someone may have to help me on stage, but I’ll win the bet.”
