Neuromodulation of cortex: modulatory signals, computational impact, and evolutionary merit

It is generally assumed that the schema of the brain is parsimonious. Given how energetically expensive it is to keep it running, the nodes, the connections, and the circuitry of the brain, need have a function that gives the organism an edge in its evolutionary struggle. That neuromodulation affects the functioning of the brain is unequivocal. These pathways are often affected in neuropsychiatric disorders, and most psychoactive substances, be it medication or recreational, have an effect on neuromodulatory receptors. Though a computational model for what neuromodulators do through their influences on the cortex, to which they profusely project, is lacking. In this thesis, I worked towards filling that lacuna of knowledge.
The first aim was to focus on one neuromodulator and find answers to when it is released in cortex, what happens to the cortical circuits on its release, and what could be the computational purpose of this. I chose cholinergic modulation of visual cortex. This choice was motivated in part by the existence of a rich but discordant body of literature as to their role in cortex. I wanted to see if I could connect the different ideas and find the source of discrepancy. Towards this aim, I discovered when the basal forebrain cholinergic axons, the primary source of acetylcholine in visual cortex, are active, and further investigated the effect of acetylcholine in vivo on neuronal and network level response metrics with laminar resolution in visual cortex. I propose a new model, in part motivated by my results, and in part by existing literature, that speculates at a computational level on why this circuit exists.
The second aim of the thesis was to extend my findings by studying a different neuromodulatory system
in multiple cortical targets. For this, I focused on serotonergic modulation of anterior cingulate cortex and visual cortex. Signals from serotonergic cortical projections have never been reported to the best of my knowledge, and this is the first study of its kind. I characterize the serotonergic signals and test an idea as to its computational role that was proposed in zebrafish.
The third aim of the thesis was to characterize a genetically labelled subset of neurons in visual cortex in
layer 2/3, and test if they are a non-random subset and fit in a tangible way in the computational model of predictive processing. The motivation here was to test if we could find a genetic handle on the positive or the negative prediction error population in layer 2/3. This would be vital to further test the predictive processing model and would be instrumental in discovering the internal representation neurons.
The fourth aim of this thesis was to test for hemodynamic signals in two-photon microscopy. I wanted to check if these are substantial to warrant attention, if we could catch them analytically by conventional means, and how uniform they are across cortical regions and layers. This is a cautionary tale that would probably inform the field.