Communication-through-Coherence or Cohernece-through-Communication

The traditional belief in neuroscience was that if two brain regions communicate, their activity will be coherent, meaning one region's activity can be predicted from the other. Specifically, coherence is a mathematical measure that refers to the synchronization or correlation between neural signals or oscillations at a specific frequency. This measure has often been used to infer if two brain regions where connected. Over time, people noticed a increase in coherence whenever information was being propagated from one brain region (the sender) to the other brain region (the receiver). Soon it was hypothesized that coherence was necessary for communication between regions. However, the review paper "Principles of large-scale neural interactions" suggests this perception is flawed.

The new hypothesis, "Communication through Resonance," is gaining more support. A resonating neuron can listen to a sender only if it is tuned to the sender's frequency. When the sender neurons' firing frequency matches the receiver neuron's resonant frequency, information is transferred. Resonance is more robust to noise and can use noise to communicate. It also operates over a broader frequency range than "entrainment." Moreover this explains why the sender and receiver appear coherent. The receiver is actively responding to the information in a particular frequency band of the sender. As such the receiver and sender are both oscillating at that frequency, and thus have a correlation and coherence in activity, albeit slightly phase offset from one another.

Resonance can also arise in recurrent networks, especially those with steady states with complex conjugate eigenvalues. This is common in networks with feedback inhibition, such as inhibitory fast spiking PV neurons. Inhibitory PV neurons in the receiver also exhibit the highest coherence. Interneurons themselves can also be resonators.

Another hypothesis is that non-linear mechanisms can boost signals, working synergisticly with resonance. Coherence is a purely linear description of a network and misses the nonlinear nature of neuron action potentials and thresholds. Non-linear mechanisms can boost momentary synchronization in sender spike-time synchronization. When multiple sender neurons have correlated inputs, the receiver neuron passes threshold and fires, making it appear as if the sender and receiver are coherent. However, if the sender neurons are not correlated with each other, the receiver neuron will not respond to the input, and the sender and receiver will not be coherent. High correlation and high coherence may result in less information transmission, such as in tremors.

In conclusion, the traditional view that coherence between two brain regions is necessary for communication is being challenged by new hypotheses. Communication through resonance, particularly in recurrent networks and interneurons, appears to be a more robust and broader mechanism for information transfer. Additionally, non-linear mechanisms that work synergisticly with resonance can boost momentary synchronization and enhance information transmission. By understanding these alternative mechanisms, we can gain deeper insights into the complex and dynamic nature of large-scale neural interactions in the brain.

Author: Alexander White