studying live brains

I don't read a lot of neuroscience, so when I ran across an unfamiliar acronym in an article about neural development, I was intruiged and had to look it up. The acronym was DTI, Diffusion Tensor Imaging. There are now a bunch of different technologies and techniques for looking at how live human brains work. In the old days, you want to wait until people died before you could cut them open. The only way you could tell what part of the human brain did what was to wait until someone got a stroke or other brain damage, and see what they could no longer do. That was good enough for the first 2/3rd of the 20th century, but it's not good enough anymore for those crazy brain scientists. Here's a brief list of all of the techniques I know of for looking at living brains and seeing what they're doing, should you run across an unfamiliar neuroscience-related acronym in your own life...

Electroencephalograpy (EEG) - This is the old and familiar technique of putting electrodes on peoples' heads to see what parts of the brain are generating electrical currents. All you can usually see are "brain waves", which refer to synchronous firing of neurons. Most of the time, you only get widespread synchronous firing during sleep and epileptic seizures.

Event-Related Potentials (ERP) - This is a technique that uses EEG to identify particular electrical spikes that occur in response to particular stimuli. For example, 1/3rd of a second after a surprising event, you get a spike of activity called the P300. ERPs are very good at telling you when something happened in the brain, but they're terrible at telling you where.

Magnetoencephalography (MEG) - This is a somewhat newer technology, somewhat similar to EEG, where horribly expensive and tempermental superconducting sensors are used to measure the magnetic field generated by neurons firing. MEG gives great temporal resolution (like the imaging techniques below), and great temporal resolution (like EEG/ERP), but it's very expensive and difficult to use.

Computed Axial Tomography (CAT) - CAT scans are very common in medicine. All they are is a bunch of x-rays taken from different directions, which a computer puts together into a 3-D x-ray. They measure density, and are very good at finding tumors and lesions and things, but they don't say much if anything about brain activity.

Positron-Emission Tomography (PET) - PET scans involve injecting radioactive sugars into people. Parts of the brain that are active need sugars (a brain's gotta eat), so the concentration of the radioactive chemical goes up in those areas. You can then triangulate where radioactive decay is coming from and build up a map of brain activity. The spacial resolution's fairly good, but the temporal resolution's not very good, and it's hard to get many people to volunteer to get radioactive chemicals put into their body in the name of science, unless it's going to cure their cancer.

Magnetic Resonance Imaging (MRI) - This is the newish standard technique for imaging brains. It involves very very large magnets spinning around your head. MRIs can detect changes in density and in the amount of oxygen in your blood. In neuroscience, people typically talk about...

Function Magnetic Resonance Imaging (fMRI) - In fMRI, two MRI images are taken and compared with each other. One is a baseline and the other is taken a few seconds after some sort of stimulus or activity. Since active parts of the brain have to eat, blood flow increases after your brain does something, and that shows up in the second image. Subtract the two, and you can see what part of the brain is more (or less) active than normal when you're, say, looking at a face. fMRI research is unbelievably popular, widespread, and trendy, despite the fact that it measures a secondary effect (blood flow, not neural activity), has pretty bad temporal resolution, and requires that experimental subjects be stuck in a loud tube for a long time to get repeated measurements.

Optical Imaging - Optical imaging is kind of a cross between EEG and MRI. Like MRI, optical imaging measures changes in blood oxygenation with activity, but instead of using magnets, they shine lights through your skull. Really. The changes in blood chemistry change the way that the light reflects off neurons in your brain, when can then be measured and processed to determine, with very good spacial accuracy, what parts of the brain are most active. However, optical imaging only works for surface layers of the brain, and although measurements can be taken continuously, like EEG, the fact that blood flow is being measured and not actual neural firing restricts optical imaging's usefulness.

Diffusion Tensor Imaging (DTI) - This is the new one to me. While MRI normally measures things related to density, DTI is a variant of MRI that measures how easily water moves at each point in the brain. The grey matter of the brain is where neuron bodies are, on the outside (mostly). The white matter is the wiring in between different areas of grey matter. White matter is made of lots of long, thin axons, which have the property that water can only move lengthwise along them. So if you have a way of measuring what direction the water is allowed to move, you can determine the orientation of the axonal wiring throughout the brain. You can then have your computer make pretty pictures that tell you exactly what part of the brain is connected to what other part!


There you are. Those are the major techniques for looking at the live human brain without cutting people's heads open.