The sources of neuroscientific knowledge

Authored by Henry Strick van Linschoten

There have always been considerable limitations on ways of acquiring knowledge about the nervous system, and especially about the human nervous system. As a result, a great deal of what is considered to be research relevant to human neurobiology in fact relies on experiments with animals. As the human nervous system in a number of ways has more similarity with the nervous system of animals than other organ systems this has been widely regarded as legitimate. However, it cannot be proven that an animal and a human will respond identically to any particular circumstance or set of events.

For a long time the main information about the brain depended on three sources:

  • autopsies
  • damage to part of the brain which enabled observation of subsequent changes in behaviour
  • electrical stimulation of brain locations

Remarkable results were gained from undertaking autopsies, and many pictures of the brain at the end of the 19th century look the same as they do now. Even more incredibly, a charting by Brodmann of the areas of the cerebral cortex based on careful microscopic study of the cortex tissue was published in 1909 and is still quoted today. The work based on both brain damage and brain stimulation established associations (correlations) and could lead to speculation, but could not produce a complete picture.

The main focus of brain research takes one of the following forms:

  • The location of a particular function of activity is determined. This requires both adequate mapping of the nervous system, as well as locating in the available structural maps where activities or functions take place.
  • It can be mapped which parts (locations) of the nervous system are connected with which other parts, and how many axons appear to connect them. Whilst this gives a rough quantitative idea of which groups of parts might work together in some systematic way, as long as it is not known what is being communicated over the connections the results have limited value.
  • Making comparisons between the left and right hand side of parts of the brain, and between the brain locations of different people, or even of the same person at different developmental stages.
  • Making size comparisons of parts of the brain, which one can then try to relate to a person’s history and genotype, i.e. genetic and environmental influences. However, “size” does not have a natural meaning.
  • For all brain processes, the dynamic over time (how and how quickly it changes) and what this might be related to, can be studied.

An overview article discussing and evaluating the relationship between psychology and biology is Miller (2010).

The new brain imaging techniques

To the old techniques already available in the 19th century, have now been added a range of imaging techniques (the first three are older than the others):

  • X-rays, but with the major limitation of the poor contrast, and the damage done by the considerable amount of radiation.
  • Angiography – an improvement on X-rays by injecting dye to give better contrast. The technique was useful for picturing blood vessels.
  • Electroencephalography (EEG), a recording of the global electrical activity (potentials) of the brain, showing up brain waves from a number of measuring positions on the skull. It is useful for investigating sleep stages and types, and for distinguishing types of seizures and seizure-like activity. It is a main source of ideas about left to right hemisphere differences.
  • Computerised tomography (CT; sometimes also CAT, with A for axial), using narrow X-ray beams, specialised detectors, and movement between the detectors and the body, thus allowing much more detailed pictures, albeit still with a resolution of several millimetres, i.e. insufficient to distinguish individual (neuron) cells.
  • Magnetic Resonance Imaging (MRI), which allows more versatility and for images to be made with a better resolution than CT, at less than a millimetre. It requires contrast agent to be used, and while it is generally regarded as much safer than CT, it is not completely safe. It has a number of variants. The basic method for making images helps to clarify physical structure, and is sometimes designated as sMRI – structural MRI.
  • Functional MRI (fMRI) does not measure neurons or neuronal function directly, but blood flow or the state of haemoglobin in the blood, which is taken as a proxy for the local activity level. It probably mainly reflects synaptic input and local processing, not neuronal impulses. As the name says, it tries to establish function and activity levels throughout the brain. The spatial resolution is a few mm, and the time resolution a few seconds, which is better than PET (Positron Emission Tomography) and SPECT (Single Photon Emission Computerised Tomography).
  • Dyadic functional MRI (dfMRI), where recently experimentally two people have been scanned with fMRI in the same machine, and lying in the machine in parallel, with the possibility of visual contact. This is bound to develop further, and is important for developing ideas in social neuroscience.
  • Diffusion MRI (dMRI) or Diffusion Tensor Imaging (DTI) is another variant of MRI. It is especially effective for images of axon pathways.
  • Positron Emission Tomography (PET) measures brain activity (rather than structure) by injecting radioactive chemicals in the blood and measuring the positron emission. Apart from the radioactive load on the subject, the measurements produced usually need large amounts of computer processing.
  • Single Photon Emission Computerised Tomography (SPECT) is simpler and lower-cost than PET, and gives images of poorer quality. It measures emitted gamma-ray radiation of the radioactive substance with a gamma camera.
  • Magnetoencephalography (MEG) measures brain activity though special ultra sensitive detectors measuring minute amounts of magnetic activity. Its time resolution is especially good, so it can measure rapidly shifting cortical activity.
  • Magnetic Source Imaging (MSI) is a combination of MEG and MRI.

When considering these techniques it is important also to be aware of their major limitations:

  • They either measure structure, which is a refined technique of making static images, or activity.
  • Measuring activity remains indirect, mostly via measuring blood flow.
  • There is a trade-off between the spatial and the temporal resolution: if the images are more precise as regards location, they cannot detect quick changes, and vice versa.
  • Many of the techniques are used in conjunction with substantial amounts of computer processing of the raw data. Some of this processing is probabilistic, smoothes out, and averages, and there is controversy as to whether what it presents as output is still in any way close enough to what one would have wished to observe directly.
  • Apart from the computer processing, many images used more widely as “results” are in fact the average of a number of subjects, the average of one person scanned repeatedly, or represent the difference between measurements taken during stimulation or at rest. Furthermore, since the brain is never inactive this also introduces an element of fuzziness.
  • The colours usually shown in pictures do not correspond to what is observed; they have been added by the computer as part of the photo-processing.
  • None of the techniques can measure activity at an individual neuronal, synaptic or molecular level, or can measure the totality of the neurotransmitter activity crossing a particular synapse.

A summary of the state of neuroimaging techniques can be found in Miller et al. (2007).


Miller, G.A. (2010). Mistreating psychology in the decades of the brain. Perspectives on Psychological Science, 5: 716-743.

Miller, G.A., Elbert, T., Sutton, B.P. & Heller, W. (2007). Innovative clinical assessment technologies: Challenges and opportunities in neuroimaging. Psychological Assessment, 19: 58-73.