I'm new in biology. could you please let me know if the following 2 assumptions are correct

A nerve cell in the CNS along with its loop back neurons (Collaterals) is a memory unit. It will trigger per given spikes pattern (example – from the PNS). The said pattern should be “close enough” to enable the memory unit to resonate.

A memory unit is a cyclic sub-graph with its “root neuron”. An inhibitory neuron eliminates multi resonance of the memory unit.

  • 1
    $\begingroup$ These things could be correct in a very specific implementation of biologically-inspired neurons for an artificial neural network, and might be reasonable assumptions to taking a graph-theory approach to understanding biological neural networks, but they don't capture the known subtleties in our actual biological understanding of how brain networks work. $\endgroup$
    – Bryan Krause
    Nov 28, 2016 at 16:23
  • $\begingroup$ For example, "memories" as abstract concepts are held in populations, not single cells, though single cells can have "some memory" - pattern completion can be performed by networks, as well. Neurons may resonate in some contexts but resonance is not necessary for recall or encoding of memories. Inhibition has many roles: it can prevent unwanted resonance (for example, seizures), help discriminate similar stimuli, control gain, suppress expected outcomes to highly unexpected events, etc. $\endgroup$
    – Bryan Krause
    Nov 28, 2016 at 16:25
  • $\begingroup$ Thanks Bryan. Let me ask you, given arbitrary episode how does the brain determine its start & end ? or when does brain conclude "I do not know" / "i do not remember" / "I do not recognize" ? $\endgroup$
    – BNR
    Nov 29, 2016 at 9:27
  • $\begingroup$ We don't know :) There are many studies that use psychophysical tasks to begin to answer some of these questions. Some have shown that estimates of confidence can be derived from activity in the prefrontal cortex - answers like "I don't know" require a bit of metacognition so it is somewhat difficult to study. It's actually quite a big deal to be able to recognize when you know and don't know, and one could imagine a fully functional animal/robot without those abilities. $\endgroup$
    – Bryan Krause
    Nov 29, 2016 at 12:41
  • $\begingroup$ i assume that the answer to this question may address my question (though it's chicken - egg issue) $\endgroup$
    – BNR
    Nov 30, 2016 at 16:06

1 Answer 1


The concepts you raise in your question are not founded in neurobiology. We do not know what the biological correlates of a "unit of memory" are in the central nervous system.

Although oscillations are widespread in the brain, the relationship between oscillations and memory storage and recall is a subject of active research. The way you state this process in your question is not justified by any experimental evidence that I am aware of.

Memories do not have a "root neuron" — as I wrote above, we do not know how memories relate to neurons. Inhibitory neurons simply provide inhibitory input currents to their post-synaptic targets. Any relationship with "resonance" or oscillations relies on explicit assumptions about emergent network properties, and these are currently unfounded.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .