From physiological principles to computational models of the cortex

Acceptably aware during general anaesthesia: 'dysanaesthesia'--the uncoupling of perception from sensory inputs

This review makes the case for 'dysanaesthesia', a term encompassing states of mind that can arise in the course of anaesthesia during surgery, characterised by an uncoupling of sensation and perceptual experience. This is reflected in a macroscopic, functional model of anaesthetically-relevant consciousness. Patients in this state can be aware of events but in a neutral way, not in pain, sometimes personally dissociated from the experiences. This makes events associated with surgery peripheral to their whole experience, such that recall is less likely and if it exists, makes any spontaneous report of awareness unlikely. This state of perception-sensation uncoupling is therefore broadly acceptable (a minimum requirement for acceptable anaesthesia) but since it is likely a dose-related phenomenon, may also represent a precursor for awareness with adverse recall. This hypothesis uniquely explains the often inconsistent responses seen during the experimental paradigm of the 'isolated forearm technique', wherein apparently anaesthetised patients exhibit a positive motor response to verbal command, but no spontaneous movement to surgery. The hypothesis can also explain the relatively high incidence of positive response to relatively direct questions for recall (e.g., using the Brice questionnaire; ∼1:500; the vast majority of these being neutral reports) versus the very low incidence of spontaneous reports of awareness (∼1:15,000; a higher proportion of these being adverse recollections). The hypothesis is consistent with relevant notions from philosophical discussions of consciousness, and neuroscientific evidence. Dysanaesthesia has important implications for research and also for the development of appropriate monitoring.

Cortical dynamics during naturalistic sensory stimulations: experiments and models

We report the results of our experimental and theoretical investigations of the neural response dynamics in primary visual cortex (V1) during naturalistic visual stimulation. We recorded Local Field Potentials (LFPs) and spiking activity from V1 of anaesthetized macaques during binocular presentation of Hollywood color movies. We analyzed these recordings with information theoretic methods, and found that visual information was encoded mainly by two bands of LFP responses: the network fluctuations measured by the phase and power of low-frequency (less than 12 Hz) LFPs; and fast gamma-range (50-100 Hz) oscillations. Both the power and phase of low frequency LFPs carried information largely complementary to that carried by spikes, whereas gamma range oscillations carried information largely redundant to that of spikes. To interpret these results within a quantitative theoretical framework, we then simulated a sparsely connected recurrent network of excitatory and inhibitory neurons receiving slowly varying naturalistic inputs, and we determined how the LFPs generated by the network encoded information about the inputs. We found that this simulated recurrent network reproduced well the experimentally observed dependency of LFP information upon frequency. This network encoded the overall strength of the input into the power of gamma-range oscillations generated by inhibitory-excitatory neural interactions, and encoded slow variations in the input by entraining the network LFP at the corresponding frequency. This dynamical behavior accounted quantitatively for the independent information carried by high and low frequency LFPs, and for the experimentally observed cross-frequency coupling between phase of slow LFPs and the power of gamma LFPs. We also present new results showing that the model's dynamics also accounted for the extra visual information that the low-frequency LFP phase of spike firing carries beyond that carried by spike rates. Overall, our results suggest biological mechanisms by which cortex can multiplex information about naturalistic sensory environments.