This is a brief return to brain waves. Neural oscillations are low-frequency synchronized activity of many neurons, from the low single Hz to 50 or 100 Hz. They do not broadcast information themselves, as a radio station might. Rather, they seem to represent coalitions of neurons being ganged together into a temporary unified process. They seem to underlie attention, integration of different sensory and cognitive modes, and perhaps subjective "binding". A recent paper provides a rare well-written presentation of the field, along with a critical approach to correlations between waves coming from different places in the brain. They also find that strength of oscillatory coupling correlates with cognitive performance.
The issue they were trying to approach was the validity of cross-frequency coupling, where a rhythm at one frequency is phase-coupled with one at a higher frequency, at some integer difference in frequency, like 1:2. (In other words, harmony was happening). Such entrainment would allow fundamentally different cognitive processes to relate to each other. They study two different types of correlation- the straight frequency coupling as above, and a phase-amplitude coupling where the amplitudes of a higher frequency oscillation are shaped to resemble a lower frequency wave. This resembles AM radio, where the high-frequency radio signal "carries" a sound signal encoded in its rising and falling amplitudes, though its frequency is completely stable. This latter form of coupling was more difficult to find, analyze, and in the end failed to have significant functional consequences, at least in this initial work.
 |
| Cartoons of cross-frequency couplings (CFC, aka harmonies) that were investigated. |
The authors' first goal was to isolate gold-standard couplings, whose participating waves come from different locations in the brain, and do not (respectively) resemble contaminating similar waves inherent in the complementary location. After isolating plenty of such cases, they then asked where such phenomena tend to take place, and do they correlate with function, like performance on tests. They used resting brains, instead of any particular task setting. This makes the study more reproducible and comparable to others, but obviously fails to offer much insight into waves as they are (if they are) used for critical task performance. Resting brains have an ongoing hum of activity, including a well-known network of waves and frequency couplings called the default mode network. Going past the authors' statistical analysis of maximally valid correlations, they found a variety of "hubs" of cross-frequency coupling, which had an interesting nature:
"In α:β and α:γ CFS, the α LF hubs were observed in PFC and medial regions that belong to the default mode network [29] or to control and salience networks in the functional parcellation based on fMRI BOLD signal fluctuations [94–96]. This is line with many previous studies that have found α oscillations in these regions to be correlated with attentional and executive functions [14–19]. In contrast, the β and γ HF hubs were found in more posterior regions such as the SM region and the occipital and temporal cortices, where β and γ oscillations are often associated with sensory processing"
Alpha, beta, and gamma are different frequency bands of neural oscillations. CFS stands for cross-frequency coupling. LF stands for the low frequency partner of the coupling, while HF stands for the high frequency partner or source. PFC stands for the prefrontal cortex, which seems to be a locus of relatively low frequency brain waves, while the sensori-motor (SM) regions are loci of higher-frequency activity. This is interesting, as our brains are generally filters that gather lots of information (high frequency) which is then winnowed down and characterized into more abstract, efficient representations, which can operate at lower frequency.
And does more correlation in the resting state mean better brain performance when doing tasks? These authors claim that yes, this is the case:
"CFS between θ–α with β–γ oscillations (θ–α: β–γ CFS) and CFS between β and γ oscillations (β:γ CFS) showed significant positive correlations with scores on Trail-Making Tests (TMTs), which measure visual attention, speed of processing, and central executive functions, as well as with Zoo Map Tests, which measure planning ability (p < 0.05, Spearman rank correlation test, Fig 8). Intriguingly, negative correlations with the test scores were observed for CFS of α and β oscillations with higher frequencies (α–β:γ) and for γ:Hγ CFS in the Digits Tests measuring WM performance.
...
These results suggest that in a trait-like manner, individual RS CFC brain dynamics are predictive of the variability in behavioral performance in separately measured tasks, which supports the notion that CFC plays a key functional role in the integration of spectrally distributed brain dynamics to support high-level cognitive functions."
RS refers to the resting state, of the subjects and their brains.
It is exciting to see this kind of work being done, gaining insight into information processing in the brain. It is an opaque, alien organ. Though it houses our most intimate thoughts and feelings, how it does so remains one of the great tasks of humanity to figure out.
