What makes it 'not psychic' in my view is that it has nothing to with extrasensory perception but normal physical principles we already understand. If magical forces were involved instead, Id have no problem viewing this achievement as 'psychic'.

Although this might be useful for more than fortune telling or the like, so maybe not. Votes: +3

Seconded. What matters is not the level of detail but the method. Even if the machine managed to print out *exactly* what the subject thought/dreamed etc, that would not count as "psychic" (i.e. supernatural) because after all the machine is physically reading brain activity. It'd be psychic if it could do this without the MRI. It is also not a "discovery". These lads will have done a lot of research and programming. "Invention" would be a better description. Better still, it is an experiment to verify all of the scientific groundwork done so far on visual processing in the brain. Science in action.

I doubt if the researchers involved would use terms such as "psychic" other than tongue in cheek. Clearly, some journalism has placed itself between us and the actual story. Votes: +11

As a researcher in this field, I despair of headlines like this. fMRI is an extremely useful tool for specific questions in cognitive neuroscience, but it suffers both at the hands of uneducated journalists all too willing to go beyond the data and make fanciful claims which the study cannot hope to justify, and also from over zealous researchers happy to bypass the requirement for careful understanding of the relationship between physiological response and cognition, before leaping to experiments which are unconstrained by theoretical models, such as the neural correlates of ... [add your favorite human characteristic here, but can include such amorphous things as a joke, orgasm, etc etc). Without knowing further detail until the paper is peer-reviewed, I would predict that this story probably represents the former. Unfortunately, the concerning frequency of such nonsense gives the field a bad name, and as a result, is often derided as neophrenology. Used appropriately, fMRI is a seminal technology, but it is often neither used nor reported appropriately, as scientists and journalists alike become seduced by the brain images it produces, which are actually only statistical maps, that is, not real images like x-ray, CT or structural MRI for example.

fMRI detects the blood-oxygen level dependent (BOLD) response; that is, blood containing (oxygenated) or lacking (deoxygenated) oxygen have different magnetic properties. Since neural activity is energy-dependent, the flow of blood to a stimulated brain region will show a mismatch between oxy- and deoxygenated blood; a comparison of task A versus task B will indicate regions in which a relative change in this ratio has occurred, and thereby serve as a proxy to brain activity. A and B are compared repeatedly, and the image represents the statistical likelihood of observing any difference by chance. The image is therefore a statistical map, not a real image reconstruction: the images do not visualize brain activity, like an x-ray visualizes bone; the 'blobs' are merely reflect areas of statistical significance, nothing more.

I explain this to illustrate that this is quantatative data - there is absolutely no qualitative information content in fMRI images. One can identify areas in visual cortex which show greater response to colour, shape, orientation, movement, etc, but it is completely beyond this technology to reconstruct images that the viewer is experiencing. It's difficult to say, given the lack of detail, but I imagine all that the researchers have done here is to produce a computational model of the likely activity that would be seen if viewing a collection of clips from YouTube (presumably based on the visual content such as speed of movement on screen, level of contrast, colour and orientation of predominant images, etc, etc), and then correlated this with the activity in the subject's fMRI image. The interpretation would then be that given pattern X of fMRI response, the most likely clip that would reproduce such activity is clip Y. If so, it's quite interesting insofar as video clips can be digitized and represented as a vector condensed across a number of perceptual parameters, and that this correlates with fMRI data. However, the devil is in the detail, and we'd need to see the guts of the analysis to determine the validity of this correlation (since this can be area of weak methodology in the field). But remember, it is no more than a correlation, there is no visual information to be derived from the image, and this is a million miles from reconstructing the perceived image from fMRI data. The journalists have undoubtedly overstepped the data, as happens so frequently with fMRI images.

For anyone interested, one of the authors published a paper detailing the method here:

(http://www.ncbi.nlm.nih.gov/pu...dinalpos=1)

I haven't read the paper, but it's important to note that their model uses only 3 components: a visual component (this would identify, at a gross level the types of visual stimuli, gradient, colour, contrast, movement, etc), a semantic component (not sure without further detail, but things may get very speculative here), and a component incorporating data on the structure and semantic content of natural images. Their task will have been to use this model to correlate fMRI data with a predefined set of images, again not to reconstruct the actual image itself, which would be impossible.

As a cognitive neuroscientist and fMRI researcher, it pains me to read such articles - if it continues, fMRI data will come to be regarded with contempt by the neuroscience community, if we're not there already, and lay sceptics will quite rightly regard the data with a level of suspicion currently reserved for devices such as L Ron Hubbard’s e-meter! Votes: +40

One important thing to note is that the computer has to be trained on the specific person whose brain you try to "read". It won't work on anyone out-of-the-box.
You need to find out how their brain responds to a given set of videos, and then you use the correlation you find to reverse the process: you find out how the brain reacts to a new video, and return as result a weighted average of the videos that give the most similar responses.
It is an interesting experiment, but I doubt you'll get a decent resolution any time soon, if ever (i.e. a clear recognizable image of what someone is seeing/thinking). Votes: +2

Last I heard, the closest anybody's gotten to a semantics of brain activity had to do with phase locking between the waves in different sections of the brain to register concepts that are momentarily conceptually corrolated. That sounds like it requires galvanometry, not O2 study. What kinds of thing are best suited for fMRI study? Votes: +0

In case anyone is interested, here is a New Scientist article on this "mind reading" software, and here is their article on an earlier experiment using still images from the same people. Votes: +3

Last I heard, the closest anybody's gotten to a semantics of brain activity had to do with phase locking between the waves in different sections of the brain to register concepts that are momentarily conceptually corrolated. That sounds like it requires galvanometry, not O2 study. What kinds of thing are best suited for fMRI study?

'Phase'locking' in this context simply refers to greater activity in task A (e.g. viewing semantically related words) compared to task B (viewing semantically unrelated words). A and B are presented periodically, either in blocks or as discrete trials, and you fit the hemodynamic response to the phase of the task presentation. fMRI is perfectly appropriate for such cognitive manipulations, provided you have some theoretical model, based on other non-fMRI data, that accounts for the cognitive difference between task A and B. The assumption in this example would be that the cognitive difference between A and B is computing the semantic distance between words (since all other requirements are notionally kept constant between the tasks, and therefore cancel each other out); relating the fMRI data to this cognitive process rests on the validity of that assumption. Votes: +5

Thanks Gazcam, it's good to hear an expert's view on this sort of story, especially with the kind of in-depth information that such sensationalist headlines always leave out Votes: +5

Even without one of those silly machines:

Entering a family restaurant and finding a coin operated psychic computer along side a coin operated biorhythm machine along side a coin operated astrology machine along side a coin operated love meter right near the checkout!

Votes: +0

@gazcam, I shouldn't worry too much about serious people suddenly deriding fMRI equipment. I've seen no end of inane "fMRI studies", but I've never blamed the machine for that. One that springs to mind is a study where heterosexual and pedophile subjects were shown naked pictures of people of various ages and genders. And what was the Great Discovery? That the same brain areas that lit up for heterosexuals viewing naked women lit up in pedophile subjects viewing naked children. Yep, they needed an MRI scanner for that.

In my own neck of the woods (electronics) people also like to use fancy gear to collect gobs of meaningless data, but there too the kit is not the cause of its being misused. Votes: +1

Here's a nice video lecture of a journalist summarising, and rightly ridiculing, some of the uses and abuses of fMRI: http://www.themonthly.com.au/n...-scan-1899

She covers the embarressment to the field that was evident in the NY op-ed piece http://www.nytimes.com/2007/11....html?_r=1, as well as the follow-up vehement objection by other scientists.

She also touches on a really fascinating study, which is likely to be of interest to all sceptics, and indeed also to those woo-woo proponents who would seek to promote their nonsense, and validate their message by accompanying it with images of brain scans. The study found that bogus explanations which included reference to imaging studies and functional neuroanatomy were rated as true, compared to when no such information was provided, and this included not only the lay public, but also neuroscience students. Scientologists, homeopaths and astrologers I'm sure all took note of that piece of research, and began applying for research grants for fMRI!!

I can only reiterate, there is much to be gained when this technology, but only when it is appropriately harnassed by scientists who are not blinded by the blobs, and reported by journalists who do not strip the data from the story, otherwise, both reduce it to something meaningless and deceiving. Votes: +5

One that springs to mind is a study where heterosexual and paedophile subjects were shown naked pictures of people of various ages and genders. And what was the Great Discovery? That the same brain areas that lit up for heterosexuals viewing naked women lit up in pedophile subjects viewing naked children. Yep, they needed an MRI scanner for that.

Hehe - as a young student, looking for a PhD area to build a career in, I was offered a project on precisely this. There is a sensible idea behind it: currently, in clinical studies, a penile plethysmograph is used to detect male sexual arousal, and this can be used to provide a physiological indication of deviant sexual desire in paedophiles where the subject would otherwise be motivated to deny such responses verbally. The problem with this method is, as any male of the species knows that even whilst viewing the image, attention can be focussed away from the image (muddy fields, muddy fields, muddy fields...), to avoid the erectile response (that one works for me!!). The supposition was, that perhaps this could be avoided with fMRI, since either the normal arousal response would be detected in the brain to inappropriate images, or the deliberate attempt to refocus away from the stimulus would also have a neural signature, which would be unavoidable on behalf of the subject, and could therefore be used to identify those who were denying deviant sexual urges, and thereby potentially a threat to children if released from mental health units. If viable, this would meet the need to replace current risk assessment, for which you might as well turn to the star charts! However, this is a good example of where the intended application outpaces the theory. The variability in brain activity between individuals is substantial, and 'interpreting' one pattern as indicative of any particular emotional or cognitive state, without appropriate theoretical and empirical constraints, is pure speculation. The implication would be, that an fMRI image would convict a person of crimes they may commit in the future, without a full understanding and appreciation of the technology providing the 'evidence'. Needless to say, I didn't undertake that particular project, and I hope, used my interest in neuroscience and fMRI for better ends. Time will tell... Votes: +8

Mr. Wagg wrote: "It's not recording what a person is thinking, but specifically what a person is watching." If I'd written this item, I'd have used "experiencing" rather than "watching." I find this phenomenon VERY interesting. And I want to know more...! Votes: +0

They put Sylvia and Uri in one of them there mind reading contraptions.

All the images were of rubes being separated from their money. Votes: +4