blog

Can a human perform as well as a machine? Usually the question is posed the other way round, and depressingly frequently answered with “much better.”

But a curious study in Anesthesiology suggests that with a bit of training, anesthesiologists are just as capable of interpreting the output, in real time, of machines that keep an eye (as it were) on patients under anesthesia as machines designed to do just that. In other words, they could be trained to interpret the so-called bispectral index used to assess depth of sedation during general anesthesia ^{(10.1097/ALN.0b013e31820e7c5c)}.

The authors used an “inverse Turing test” to compare the performance of trained versus untrained anesthesiologists with a commercial monitor. As F1000 Member James Szocik puts it, “The reverse Turing test checks if a human can act like a machine, in this instance, can an anesthesiologist act like a BIS monitor? The answer is ‘Yes’. ” The paper suggests that formal education in interpreting electroencephalographs would be beneficial for patient care, and that “nonproprietary electroencephalographic measures in conjunction with clinical context may be as informative as commercially available depth-of-anesthesia monitors.” In other words, 45 minutes of training could be cheaper than the machine that goes ‘ping’.

And the authors are to be applauded for the dedication in their paper, which will warm the hearts of geeks everywhere:

The authors dedicated this study to Alan Turing Ph.D., O.B.E., cryptographer, mathematician, logician, philosopher, and founder of computer science. He contributed much to science and to humanity.