Is this really mind-reading?

http://www.guardian.co.uk/commentisfree/2012/feb/01/is-this-really-mind-reading

Version 0 of 1.

You've probably seen the headlines by now. The Daily Telegraph ran a story headlined "Mind reading device could become reality", the Sydney Morning Herald hailed the "breakthrough in mind reading technology", and popular gadget blog Gizmodo told us that "scientists can now actually read your mind". Even the Guardian got in on the act, with "Mind-reading program translates brain activity into words", while New Scientist went all out with Telepathy machine reconstructs speech from brainwaves. The headlines refer to a study published yesterday in the journal PLoS Biology, led by Brian Pasley of the University of California, Berkeley. Pasley and his colleagues had a rare opportunity to record human brain activity directly, from conscious patients undergoing evaluation before having brain tumours or abnormal, seizure-causing tissue surgically removed.

During this evaluation, the researchers placed electrode arrays on to the surface of the patients' brains, over a region called the posterior superior temporal gyrus (pSTG), which is known to be involved in speech comprehension. They played pre-recorded words and sentences to the patients, and recorded the responses of cells in the pSTG to the recordings. This enabled them to identify the key features of spoken words that the cells respond to, such as the rate of syllables and the changes in volume over time. Using specially developed algorithms, the researchers were able to reconstruct the responses and thus play back the sounds accurately enough to recognise individual words.

Most studies that garner "mind reading" headlines use functional magnetic resonance imaging (fMRI, or brain-scanning), to record the brain's responses to certain stimuli. Typically, subjects are shown a limited set of stimuli, such as pictures, and their brains are scanned to record activity in the visual cortex associated with each one. They are then shown the same set of images again, and the responses are compared to the pre-determined activity patterns to predict which one they are viewing at any given time.

Recent advances in these techniques now enable researchers to do something even more remarkable – they can decode brain activity to accurately reconstruct novel images that the subjects have never seen before. And last year, another team of researchers from Berkeley extended this approach to reconstruct novel pictures from brain activity. This is possible because the primary visual cortex contains groups of neurons that respond in stereotypical ways to specific features of an image, such as contrast and the orientation of edges.

The latest study is unique because the neural activity that was reconstructed was recorded directly from the brain using electrodes, and not by brain scanning. But the underlying principles are basically the same. We should not underestimate the achievements of the researchers involved. All of the studies mentioned here are extremely challenging from a technical point of view, and are important landmarks in how these technologies are being used. And some of them also give us useful information about how certain parts of the brain work.

But do they constitute mind-reading? My answer would have to be "no". Our thoughts are a continuous and extremely rich stream of consciousness containing multiple representations of sights, sounds, feelings, memories and more, and the corresponding brain activity is infinitely more complex than that associated with words or images. Despite these remarkable achievements, we are nowhere near being able to decode and reconstruct such complex neural activity. It may, in fact, never be possible – your thoughts will remain private for a long time yet.

• Follow Comment is free on Twitter @commentisfree