Introductory Remarks

Neuronal function underlying behaviour and cognition is organized on many different levels: from genes and molecules to cells to neuronal networks and large scale neuronal systems involving distinct brain nuclei and regions. Only the knowledge how these levels of organizations interact to generate emergent properties on the next higher level lets us fully appreciate the brain´s functionality, and allows us to use this understanding for applications in medicine and technology. The idea that complex brain function or neurological disease can be directly and causally related to whether individual genes are intact or deficient is not realistic for most of the brain functions or dysfunctions studied. The reason is that the diseases where it is theoretically possible to ‘repair’ a single gene and thereby heal the patient constitute only a tiny minority within the entire spectrum of known brain diseases. By the same token, no complex brain function has been found that is caused by one single gene. The reason is that most neuronal (dys-)functions are not ‘caused’ by genes in the strict sense. It is rather the case that complex networks of genes ‘predispose’ the brain to function or dysfunction. Theoretically, only the knowledge of the exact composition of a vast pool of genes as well as the contributions of environmental factors allows one to predict exactly the expression of neuronal (dys-)function. Importantly, such predisposing sets of genes affect whole networks of molecules and cells on the higher levels of organization in ways as yet unknown.

Therefore, in order to understand thoroughly what is going on when complex brain function is generated, neuroscientists must prepare for the laborious and arduous task of finding out how function is generated on each level of organization and how it is transferred to the next higher. This rests on the insight originally expressed by David Marr and Werner Reichhardt and is now the mission of the CIN.

Research on each level of organization has given rise to methods that are focused on the typical scale of that level and its particular requirements. The following three panels show the spatial (ordinate) and temporal (abscissa) resolution of the neurobiological tools that are most commonly used. The degree of resolution that is roughly feasible with each of the tools is demarcated by the coloured boxes. Note the logarithmic scale.

Among the methods that are applicable for study of the human brain, the electrophysiological methods EEG and MEG have excellent temporal but rather poor spatial resolution. Electrical activity up to 100 Hz can be resolved (millisecond resolution) but the spatial specificity is at the level of the whole brain or sometimes of the cortical lobe. Modern imaging techniques are better at locating brain activity, but suffer severe drawbacks in terms of temporal resolution. Neither functional Magnetic Resonance Imaging (fMRI) nor Positron Emission Tomography (PET) report neuronal activity on a scale better than a few seconds (cognition and behaviour). Similar constraints are also valid for lesions, which are the most traditional means of studying human brain function. Lesions due to disease (strokes or tumours) usually affect large brain areas and are only rarely small and circumscribed enough to allow conclusions to be drawn about the function of a single defined brain structure or neuronal pathway.

Electrophysiology using microelectrodes is in many ways still the gold standard of modern neurophysiological research. It has excellent spatial and temporal resolution in the millisecond and micrometer range. Intracellular techniques, such as patch clamp recordings, can yield data from subcellular compartments such as dendrites and axons (cells). Extracellular recording yields information about single action potentials from single neurons (single unit). Field potentials sample from small networks confined to the sub-millimeter range. Around the same spatial range of neuronal tissue is activated by electrical micro-stimulation (networks).