To construct spatio-temporal maps of brain activity, it is important to establish how neural activity is recorded. From the foundation I gave in previous sections, a description of neural activity includes action potentials which can be organised in time by slower neuronal oscillations.
Hans Berger, close to one hundred years ago, showed that electrical oscillations from the cortical surface of the brain detected by electro-encephalograms (EEG) did not cease even when the subject was not performing any specific task1. These electrical oscillations, produced by neurons, occur throughout the central nervous system in higher vertebrates including the spinal cord, subcortical deep brain nuclei, and cerebral cortex, generating what are known as ‘brain waves’2.
Following historical traditions, the vast majority of the studies examining EEG oscillations rely on canonical frequency bands, which are approximately defined as: infraslow (<0.1 Hz), delta (1–4 Hz), theta (4–8 Hz), alpha (8–12 Hz), beta (12–30 Hz), low gamma (30–60 Hz), high-frequency activity (60–250 Hz) and fast ripples (200–400 Hz). The electrical brain waves in the human cerebral cortex have been classified into broad frequency bands including ultraslow (<1Hz) delta (0.5–4 Hz), theta (4–7 Hz), mu (8-12 Hz), alpha (8–12 Hz), beta (13–30 Hz), and gamma (31–100 Hz) bands. Infra-slow oscillations (<0.1 Hz) have also been observed in multiple subcortical brain regions, including the hippocampus, basal ganglia, and locus coeruleus in humans and rodents.
Although conceptually there is no reason to separate such frequency bands as separate entities except for the power (prevalence) distribution with peaks at those frequencies in different conditions, some broad correlations between frequency bands and certain brain functions have been investigated. For example “Delta waves (less than 4 Hz) are characteristic of deep sleep and coma. Theta waves (4–8 Hz) are characteristic of emotions, memory and other limbic activity. Alpha waves (8–12 Hz) are characteristic of awakened states in occipital and frontal cortex when active processing is not occurring. Beta waves (13–30 Hz) are found in frontal and prefrontal cortex when one is alert and actively processing. Finally, gamma waves (greater than 30 Hz) are found in prefrontal and parietal association areas when the brain is working hard”3. The EEC and magnetoencephalography (MEG) gives excellent time discrimination of electrical brain events (ms range) but poor spatial information.
More recently functional magnetic resonance imaging (fMRI) methods have been introduced. These are based on the fact that the local level of oxygen changes during brain activity. When neurons become active, local blood flow to those brain regions increases, and oxygenated blood displaces oxygen-deoxygenated blood. Deoxygenated hemoglobin is more magnetic (paramagnetic) than oxygenated haemoglobin (Hb), which is virtually resistant to magnetism (diamagnetic). This difference can be mapped to show which neurons are active at any particular time. The haemodynamic lag (the amount of time it takes for local blood oxygen levels to increase) takes seconds. Consequently, the temporal resolution of fMRI is limited to several (4 to 6) seconds. More recently, faster acquisitions have emerged enabling major improvements in spatial and/or temporal resolution; for example, acquiring data with 2 mm spatial resolution in less than 1 s.
Reentrant connections between different cortical areas, together with local negative feedback loops, give rise to exceedingly complex dynamics that are characterized by oscillations in a broad range of frequencies, synchronization of discharges, and cross-frequency coupling4.
- Hans Berger (1929): Über des Elektrenkephalogramm des Menschen. Archiv für Psychiatrie und Nervenkrankheiten. 87, 527-580 ↩︎
- G Buzsáki & A Draguhn (2004): Neuronal oscillations in cortical networks. Science 304, 1926-1929.
↩︎ - Rex Welshon (2010): Philosophy, Neuroscience and Consciousness. Taylor & Francis. ↩︎
- W Singer (2013): Cortical dynamics revisited. Trends in Cognitive Science 17, 616-626. ↩︎