00:00 We looked at scans acquired in over 2,500 people from, and the data was acquired at
00:10 25 different locations around the world, so it was a really collaborative effort.
00:14 And having this large data set allowed us to use some new statistical techniques where
00:18 we can train a model to learn expectations for what the general population should look
00:24 like in terms of brain size.
00:26 So much like a growth chart that's used in paediatric medicine, many parents might be
00:31 familiar with taking their newborn to get things like height, weight and head circumference
00:36 measured and the doctor will compare those values with what they see in the general population
00:41 to see if the baby is growing in the right way.
00:43 We can now do a similar thing with the brain and we did this looking at the size of over
00:48 a thousand different brain regions.
00:50 And the six conditions we looked at were schizophrenia, attention deficit, hyperactivity disorder,
00:55 major depression, bipolar disorder, obsessive compulsive disorder and autism spectrum disorder.
01:01 We found that people with any one of those conditions were generally more likely to show
01:05 a deviation from population expectations, so areas where we see unusually large or small
01:10 volume estimates relative to the population.
01:14 But the specific areas in which these deviations occurred were highly variable.
01:18 So for any group of people with the same diagnosis, we didn't find an overlap in, so not more
01:24 than 7% of people showed a deviation in the same brain region.
01:28 So this tells us that any attempts to try to identify a single underlying cause of these
01:32 conditions or a one size fits all approach to treatments is only likely to be effective
01:39 for a small fraction of people.
01:42 But we did find that these deviations aggregated within common brain circuits and networks.
01:46 So this tells us that by trying to target brain networks or circuits rather than individual
01:51 areas, this might be a more fruitful way of trying to develop better treatments.
01:55 [BLANK_AUDIO]
Comments