Models and musings about them.

Theories of brain function abound, and they range from Aristotle's idea that it cools the blood to the most recent conclusions deduced from fMRI scans. Today, such ideas, theories and constructs are often called models, the best of which blend the writer's laboratory data with what he or she has culled from the experiments others report. A model can therefore be viewed as the product of a unique collection of intellectual encounters with teachers-some living, some dead-at lectures and in libraries to which have been added countless unique personal experiences during experiments performed in laboratories. This essay describes some of the models I have published, naming the places where the experiments were done and some of the teachers from whom I learned what I know.

Retino-cortical stimulus frequency-dependent gamma coupling: evidence and functional implications of oscillatory potentials.

Long-range gamma band EEG oscillations mediate information transmission between distant brain regions. Gamma band-based coupling may not be restricted to cortex-to-cortex communication but may include extracortical parts of the visual system. The retinogram and visual event-related evoked potentials exhibit time-locked, forward propagating oscillations that are candidates of gamma oscillatory coupling between the retina and the visual cortex. In this study, we tested if this gamma coupling is present as indicated by the coherence of gamma-range (70-200 Hz) oscillatory potentials (OPs) recorded simultaneously from the retina and the primary visual cortex in freely moving, adult rats. We found significant retino-cortical OP coherence in a wide range of stimulus duration (0.01-1000 msec), stimulus intensity (800-5000 mcd/mm2), interstimulus interval (10-400 msec), and stimulus frequency (0.25-25 Hz). However, at low stimulus frequencies, the OPs were time-locked, flickering light at 25 Hz entrained continuous OP coherence (steady-state response, SSR). Our results suggest that the retina and the visual cortex exhibit oscillatory coupling at high-gamma frequency with precise time locking and synchronization of information transfer from the retina to the visual cortex, similar to cortico-cortical gamma coupling. The temporal fusion of retino-cortical gamma coherence at stimulus rates of theater movies may explain the mechanism of the visual illusion of continuity. How visual perception depends on early transformations of ascending sensory information is incompletely understood. By simultaneous measurement of flash-evoked potentials in the retina and the visual cortex in awake, freely moving rats, we demonstrate for the first time that time-locked gamma oscillatory potentials exhibit stable retino-cortical synchrony across a wide range of stimulus parameters and that the temporal continuity of coherence changes with stimulus frequency according to the expected change in the visual illusion of continuity.

The human Retinal Functional Unit.

It has long been known that readers of this page will move their eyes from one fixation to the next two to four times per second. It follows from this fact that each fixation triggers a unique optic nerve volley lasting up to 300 ms that contains all the information the retina processes between fixations. Here we give such volleys a name (Retinal Functional Unit, RFU) and use human subjects and interstimulus interval (ISI) experiments to define some of their properties. We report that RFUs can be dissected into an initial fraction that reaches the cortex and a later fraction that may not, depending on the ISI between successive stimuli. During the dissection process the perceptions of the stimuli change in an orderly way, such that successive thresholds of "twoness", color, and duration are reached as a function of increasing ISI. We conclude that volleys from the tens or hundreds of thousands of active axons contained in every RFU exit the retina in a precisely determined temporal order, and add this conclusion to three others for which we have already published the supporting data. 1) The mammalian retina normally takes about 300 ms to process a visual stimulus. 2) The ca. 300 ms end product, an RFU, contains in neuronal form all the photochemical information acquired during one fixation. 3) These information-rich volleys reach the cortex with little or no change thanks to monosynaptic transfer in the thalamus.

Reflections on the conceptual origins of neuron--glia interactions.

Glial contributions to the functioning nervous system are now recognized widely. This was not always so. What follows is an account of the several mid-20th century events that led to the explicit proposition that glia and neurons collaborate in the production of behavioral responses.