Whoops...

It occured to me that in my last post I never actually mentioned the DYNAMICS of my job.....like, what I'm actually doing.



So I'm working at the Psychiatry Neuroimaging Lab (Which I already stated) primarily doing research on schizophrenia and other related psychological disorders. Okay, so thats fantastic (seriously), but how the heck do you do that? Well, its simple really. Data, Statistics, and a TON of processing speed.....wicked fast computers..... Everything at the lab is done in Linux, which is primarily driven by code. Its like windows, but when you ask it to do something, it actually does it-Crazy, I know. Learning the code was a pain, but I'm slowly starting to pick it up.



So, as many of you know Schizophrenia has a considerable amount to do with THE BRAIN (Duh). Which means theres a ton of imaging done. There are thousands of cases logged into the PNL's database (Both Normal Controls and Schizophrenics) and each one has tons of scans done to it:MRI (Both T1 and T2 images), fMRI, DTI, the list goes on, but most of these measure the activity levels in the Brains of these case studies. So once all these scans are collected, RA's at the lab "polish" them off, going through slice by slice in an MRI scan (In a program aptly named "Slicer") and "masking" them, which means filling in the holes that the scan missed. They do this to make sure the total area/volume of the brain and regions within it can be measured, and analyzed itself...



Which is where Freesurfer comes in....OH. MY. GOD. This is the COOLEST FLIPPING PROGRAM EVER!!!!! It takes all the "polished" brain scans and makes an outrageously close model of that case's brain. AND THEN LABELS AND DIVIDES THE BRAIN BY REGION!!! ITS SO COOL!!! I'll see if I can post an image or something, but IT REALLY IS AWESOME!!! By dividing up the regions, their Gray Matter volumes, Surface Area, Cortical Thickness, And a whole bunch of other really huge stats. Basicallly, this provides the bulk of the information which is analyzed in the projects. Which brings us to its trade-off.....In order to "create" a brain for freesurfer, that is, in order for freesurfer to create a model of a Case's brain, the t1 and t2 images have to be processed....for THIRTY HOURS.....So you need some crazy processors (Everythings Xeon at the PNL) and a lot of patience....

BUT WAIT, THERES MORE!!! So freesurfer gives us the info, but how do you statistically beat the crap out of it until you can recall the Gray Matter Volumes for all Forty Cases in order of region of study? WITH THE DATA MINING SPSS SOFTWARE, OF COURSE!!! This is like Excel on Steroids for all you Microsoft Office Fans out there....You put all the data into a spreadsheet just like excel, but each column acts as a variable, and you tell the program to define which variable is which, and which depends on which, and then you analyze. Independent Sample T test, Covariation, Correlation, R squared, P values, Statistical Significance.....Eat your heart out, Mr. Burbank.....

And that raps up today's rant. Tomorrow, I'll tell you all about my actual research project, and what I get to do this summer....But I think a three page wall post covers everybody's questions...