A functional magnetic resonance imaging (fMRI) – a brain research method, which measures the change in the current brain blood test person based on nuclear magnetic resonance phenomenon. These changes are due to changes of the brain activity, since the intensity of the neurons associated with cerebral blood flow: the more active work one or another region of the brain, the more blood rushes to her
<- place 8255567, / science / 2016.! /05/22_a_8255567.shtml,nm2015/v2/article/incut,incut1_link ->
That is why fMRI actively used by researchers all over the world for the study of the brain. As a rule, in the course of scientists experiment, asking the person to perform various tasks: to read the text, watch a video, listen to certain sounds, take it in your mind, play a game of chance, etc.
. at this time, the brain is scanned using a powerful magnetic field, whereby the fixed and the smallest changes in the blood flow.
Scientists emphasize that when the researchers interpret the data obtained in the course of brain scan, they do not look directly at the brain – they look at brain image, which, as it is made up of tiny cells (also called voxels). The image obtained as a result of a special software that processes the data during fMRI. It turns out that computer programs are the intermediary between the equipment and the MRI scientists. “Software – not man – scans voxels, combining them into clusters, – says one of the authors of Richard Chirgvin. – When you see the phrase “Scientists can predict the movement of your arm – it tells the brain the image”, it really is a question of images obtained with the help of software »
A team of researchers led by Anders Eklund from Linköping. University (Sweden) decided to check how much you can trust the software used today, namely, the three most popular software packages – SPM, FSL and AFNI. The scientists took the results of the brain fMRI studies 499 healthy people, who during the examination of the brain is at rest. It would be logical to assume that the results of fMRI have had to be very similar, if not identical.
Scientists assumed that the percentage of false signal recognition will not exceed 5%
. <-! place 8296565, / science / 2016/06 / 10_a_8296565.shtml, nm2015 / v2 / article / incut, incut2_link ->
The subjects were divided into groups of 20 people – researchers compared the results of the fMRI on individual parameters. Total researchers have received nearly 3 million “points of comparison”. After the work was completed, the researchers found that in fact the percentage of false signals recognition was 70%. Wine was in error operation of the software – in more detail with the conclusions of the authors can be found in their article, published in the journal PNAS
However, the worst news was not even detect the most errors in the software. providing and what present she is there for the past 15 years (by the way, the bug fixed in May 2015 – it was then that the research team of Anders Eklund received your information and start work on the publication)
This means that about 40 thousand. research, which used data from fMRI, may be incorrect.
Moreover, bugs in the software packages have led not only to an inaccurate interpreting these imaging – it came even to the point that the image indicative of the active work of the brain, while actually he was at rest
If we consider that in studies using fMRI, usually takes a small number of participants. (about 30 – this is due to the very high cost of the equipment, about $ 600 per hour) , the results of many scientific papers will be all the more doubt. Another problem is also caused by the high cost of the equipment, is that researchers simply share data between themselves: interpretation of the fMRI results, obtained by a research group, were laid in sharing so it could learn everyone
<.! - place 8148617, / science ™ / 2016/03 / 29_a_8148617.shtml, nm2015 / v2 / article / incut, incut3_link ->
Caution urged to treat this research fellow at the Center for cognitive studies and neuroeconomics, head of the psychology department of the HSE Basil Klyucharev. In his studies, he and his colleagues in the field of neuroscience, said software package uses SPM. “Indeed, there is a problem – when we study the brain by MRI, we have a huge amount of data. If the photo we see the pixels, the three-dimensional brain, and then it comes to the voxels, which can be up to 50-80 thousand, they contain information about the activity of different brain regions, “-. Klyucharev told” Gazeta.ru “
For this reason, he said,
many scientists to detect activity in a small area of the brain, analyze activity only in it, without taking into account the activity of the whole the brain.
so neuroscientists “play” with many options of software settings, looking for new statistical approaches to identify significant activity in the brain, and debates are among them on how to interpret the different experimental data . “If you study this separate area, strictly speaking, correct, taking into account the activity of the entire brain, statistically never prove that there is” – said the scientist. According to exit – Search repeatability and application of meta-analysis: take a large amount of experimental data on one topic and analyze it statistically. Basil Klyucharev also said that Article MRI critics have published another critical article , which are disadvantages of the methods used by Eklund team.
No comments:
Post a Comment