Research Methods — Basic Neuroscience and Geometry of Neuroscience

The lecture opened with a memory-test question on how smells are processed, which is collected with the other questions at the end.

Research: the reason we know anything

Definition 10.1 (The scientific method). The scientific method is an empirical and objective method of acquiring knowledge through observation, skepticism and repetition. It rests on

hypothesis testing based on deductive reasoning, and
mathematical and statistical evidence for natural and observed phenomena.

The slide’s portrait is of Ibn al-Haytham (965–1039), a polymath considered the father of modern scientific methodology because of his emphasis on experimental data and on the reproducibility of results.

Science has not always been kind to the people who moved it forward. The next slide, “It wasn’t always smooth sailing,” shows three of them: Nicolaus Copernicus (1473–1543), drawn with a globe and a chart of the heavens; Ignaz Semmelweis (1818–1865); and Sergei Korolev (1907–1966), whose picture is a police mugshot. In the lecturer’s words, they made pretty major contributions, and were thanked for them in pretty much the worst ways possible. Their stories were left as a Wikipedia dive for anyone curious.

Research today

Today, experiments are carried out and published in scientific journals (the slide shows the cover of Nature from 15 February 2001, the issue on the human genome).

All research goes through peer review before it is published.
If the science is particularly good, it later gets compiled into your textbooks.

Universities are largely research institutes.

Why do we have ethics boards regulating research?

Example 10.2 (The Tuskegee Syphilis Study (1932–1972)).

It was a study of the neurodegenerative effects of syphilis in Black men.
The men were told they were receiving free medical treatment. This was a lie.
A treatment for syphilis became available in the early 1940s, but the experiment continued and the participants were never told.
Only a leak to the press got the study shut down. By then 82% of the participants had died, more than 30% of them from syphilis.

The slide’s photograph shows a doctor drawing blood from one of the men.

This study is a major reason we have ethics boards in research today. As the lecturer put it, you can talk about the Stanford Prison Experiment and the Milgram shock study, but there were far more studies that were far more harmful, and this is one case of many.

Institutional regulation of research

Two boards oversee research.

Definition 10.3.

The IRB (Institutional Review Board, 1974) oversees all research using human subjects, or their data.
The IACUC (Institutional Animal Care and Use Committee, required in its current form since the mid-1980s) oversees all research using non-human animals.

The animal research the IACUC oversees is empirical research. It is unrelated to “animal testing” of cosmetics, which is banned under European Union law and restricted in the USA following the Animal Welfare Act.

Animal research follows the 3Rs: Replacement, Reduction and Refinement. The practices and regulations that ensure animals’ care are constantly being re-evaluated and updated based on the latest empirical findings and legislation.

Learning from neurological damage

At first, what we could learn about the brain came from people who had brain damage, and we still learn much from medical research today. In the lecturer’s words, most of what we know from neuroscience research in humans more or less hangs on waiting for disaster to strike: researchers look at what functions people with brain damage can no longer do, and try to match each lost function to the damaged part of the brain. The approach is obviously limited, but it is where the field came from.

The reasoning is deduction by dysfunction: if a region is damaged and a behavior or process is dysfunctional, that brain region must regulate it.

Definition 10.4. Modularity is the idea that specific brain regions have specific functions.

The slide’s pictures: a 19th-century portrait of a man holding a long iron rod, and a drawing of a skull with that rod driven up through it from below the cheekbone to the top of the head (Phineas Gage); two surgeons in caps and gowns studying an X-ray of a skull; and 3D reconstructions of two damaged brains, seen from the front and from below, each with a dark area of damage in one frontal lobe.

Brain imaging in history

Early methods of looking at the brain worked, as the lecturer put it, very poorly.

1880s: the human circulation balance. A non-invasive way of analyzing blood flow during emotional and intellectual activity.
1920s: ventriculography. Imaged the brain’s ventricles (its fluid-filled cavities) by injecting air into them, which shows up on an X-ray. It was painful and often led to hemorrhaging.
Angiography (introduced in 1927). Imaged the brain’s blood vessels with great clarity. The slide links it to the Nobel Prize in Physiology or Medicine: its inventor, Egas Moniz, did win that prize, in 1949, though for a different procedure, prefrontal leucotomy (lobotomy).

The slide’s pictures: a drawing of the balance, with a person lying on a long table balanced on a central pivot; a patient strapped into a large circular metal frame; and an X-ray room with a patient on the table.

MRI: magnetic resonance imaging

Magnetic resonance imaging (MRI) came along in the 1970s.

The head is placed in a strong magnetic field.
The magnetic field is disrupted by the water in the brain, and this can be translated into structural imaging of the brain’s layers.
It earned a Nobel Prize in Physiology or Medicine.

What is it useful for? This finally gave us some useful information: MRI can show the structure of the brain, but that’s it. The slide shows a scanner and an MRI of a head sliced down the middle, front to back, with the brain’s structure in fine detail.

The next slide sets an MRI beside a CT scan of a head, both sliced horizontally. The MRI shows the folds of the cortex in sharp detail, along with a large region between the skull and the brain on one side, where the surface of the brain is pushed inward. The CT shows the skull bright white but the brain itself in much less detail.

fMRI: functional magnetic resonance imaging

Functional magnetic resonance imaging (fMRI) dates from around 1990.

Definition 10.5. fMRI measures changes in blood oxygenation through the brain as a proxy for brain activity. Firing action potentials consumes oxygen carried in the blood, so a region using more blood oxygen is taken to be more active.

More precisely, fMRI tracks the changing balance of oxygenated and deoxygenated hemoglobin. As the lecturer stressed, we have to be precise about what is being measured: what these scans really measure is changes in blood oxygenation.

fMRI uses the subtraction technique: measure at baseline, then measure again during a task or stimulus, and subtract the baseline to detect the change.

The slide’s scans show three conditions, labeled left hand, right hand and both feet, each as a horizontal slice of the brain. In the top row the whole brain is color-coded by signal change (from to ); in the bottom row only the strongest changes (from to ) are shown, in yellow and red on a gray anatomical image. The two hands light up circled patches on opposite sides of the brain, and the feet light up a patch near the midline.

fNIRS: functional near-infrared spectroscopy

Functional near-infrared spectroscopy (fNIRS) is a more portable and affordable version of fMRI: it tracks blood oxygenation using near-infrared light, from a cap worn on the head (the slide shows one, wired to a recording unit and a laptop). It is limited to analyzing surface-level neurons near the skull, and it localizes activity less reliably. In the lecturer’s words, it is basically a portable, limited version of fMRI: its spatial resolution (how precisely it can tell where activity is) is very limited, and anything deeper in the brain is unreachable.

Be careful with interpretation

Example 10.6 (The dead salmon). Some researchers were concerned that the image-processing techniques used on fMRI data could let people artificially create results. To make the point, they put a dead fish from a grocery store into an fMRI machine.

The slide quotes their methods section. The subject was one mature Atlantic salmon (Salmo salar), about 18 inches long, weighing 3.8 lbs, and “not alive at the time of scanning.” The task was an open-ended mentalizing task: the salmon was shown a series of photographs of people in social situations with a specified emotional valence, and was asked to determine what emotion the person in each photo must have been experiencing. The 15 photos were presented in a block design, each for 10 seconds followed by 12 seconds of rest, for a total scan time of 5.5 minutes.

Depending on how the results were processed, neural “activity” could be erroneously detected: the slide’s scans show small red spots in the salmon’s brain and spine, on a color scale of -values from 2.5 to 4.5. As the lecturer put it, a dead salmon had a profound neural response to emotion-evoking pictures. A dead salmon they bought at a grocery store.

The study is used as a word of caution about how fMRI results are produced and interpreted: you can make these machines show you whatever you want to see, depending on how you calibrate them and handle the data. It won the Ig Nobel Prize, a prize for research that’s just a little too weird.

EEG: electroencephalography

An electroencephalogram (EEG) uses several electrodes on the scalp to measure changes in the brain’s electrical activity over time.

It can only measure very large chunks of brain, without much specificity for regions.
It is limited by the number of electrodes, and to regions of the brain close to the skull.

What is it useful for? EEG has very limited spatial resolution, but it is quite portable and gives decent broad-scale readings of neural activity with good temporal resolution (it tracks precisely when activity changes).

The slide’s figure shows a head wearing electrodes and an EEG recording of eight channels, four from the left brain and four from the right. Partway through, the left-brain traces change from small irregular waves to large, rhythmic, spiking ones: a seizure. The right-brain traces stay normal throughout.

Histology

Definition 10.7. Histology is imaging cells and proteins within slices of tissue samples. Specialized dyes and antibodies stain specific proteins or cell types, which are then imaged under a microscope.

Histology can look at specific cell types and quantify neuronal activation at a specific snapshot in time. What is it useful for? If you want to look at specific types of neurons, receptors, or cells that produce specific neurotransmitters, it is basically the only good way to do it currently. The slide’s image is a fluorescent stain of a slice of tissue: a scatter of cells marked in red, with a smaller group of neurons, and their branches, picked out in green.

Optogenetics

Definition 10.8. Optogenetics switches neurons “on” and “off” using light.

It can selectively inhibit or activate neuron populations and pathways.
It works in real time, so it is more useful than blocking neurons or receptors with drugs or with transgenic lines (animals bred with altered genes).

The slide’s illustration shows a beam of blue light shining onto a neuron’s cell body. What is it useful for? It is currently the flashiest, and probably the easiest, way to actually manipulate neuronal activity in real time.

DREADDs

Definition 10.9. DREADDs (Designer Receptors Exclusively Activated by Designer Drugs) are a “chemogenetic” alternative to optogenetics: neurons are switched on or off with a drug instead of light.

A harmless virus carries the gene for a genetically modified receptor into the brain, and the target neurons make the receptor.
Later, giving the activator drug activates or deactivates those neurons at will.

The slide’s diagram traces the steps. A piece of DNA carries two genes: one for GFP (green fluorescent protein, a fluorescent tag that makes the modified cells visible) and one for the DREADD itself (the diagram names two versions, hM3 and hM4; hM3 switches neurons on and hM4 switches them off). The DNA is packed into a viral vector, which is injected into the brain of a mouse. Later the drug CNO (clozapine-N-oxide) is given orally or by injection, and it acts only on the cells carrying the designer receptor.

DREADDs are not as instantaneous as optogenetics, but they are much cheaper.

Check yourself

Question 10.10 (Memory test). Olfactory signals are processed in several brain areas, including the amygdala, the hippocampus, the prefrontal cortex and the hypothalamus. Why is each of these brain areas important for processing olfactory information?

Show solution

From Lecture 1: the amygdala handles emotion, fear and stress, so a smell can set off an emotional or fear response. The hippocampus stores and retrieves memories, so smells become tied to memories. The prefrontal cortex handles decision-making, and its orbitofrontal part assesses reward and punishment, so it judges whether a smell is good or bad and what to do about it. The hypothalamus is the gateway to the periphery and controls the neuro-endocrine system (stress, sexual behavior, social behaviors), so smells can trigger hormonal and behavioral responses.

Question 10.11 (Exit discussion). Which technologies would be best suited to answering these research questions?

Are dopamine neurons activated by a drug?
How does seeing a disturbing image affect the amygdala?
How quickly is a tumor growing?
How is a specific neural pathway involved in a behavior or function?