Integrative function in the marmoset lateral geniculate nucleus

Abstract

The lateral geniculate nucleus (LGN) is the principal recipient of signals from the retina and is the main source of visual input to the cortex. The information the LGN sends to the visual cortex is not a simple relay of retinal signals. It is also clearly affected by inputs to the LGN from a variety of sub-cortical nuclei. The common marmoset (Calithrix jacchus) is a New- World diurnal monkey, which is becoming increasingly popular as a model for the human visual system. Since, so far, the sources of sub-cortical inputs to the LGN of marmoset have not been clearly delineated, the sources of sub-cortical inputs to the marmoset LGN and some of the putative effects those inputs might exert on LGN signals are the principal focus of this thesis. Chapter 1 provides an overall introduction to the functional anatomy of the primate LGN and its connections. Chapter 2 constitutes an anatomical study examining the organisation of brainstem, tectal and pretectal inputs to the marmoset LGN and how it compares to the organisation of those inputs in other primates and several non-primate mammalian species. A retrograde tracer (Microruby) was injected into the LGN of an anaesthetised marmoset. The visuotopically organized nuclei such as superior colliculus, its ‘satellite’ parabigeminal nucleus and pretectal nucleus of the optic tract and were all found to send projections to the LGN. We were able to establish that the collicular projection to LGN is visuotopic. Overall, the organisation of sub-cortical projections to the marmoset LGN is very similar to the organisation of sub-cortical projections to LGN found in other primate species, but differs quite substantially from those in murid rodents or eutherian carnivores. Chapter 3 investigates the shared variability (noise correlations) between the LGN neurons. Using a 32-channel array, spike activity of many single cells was recorded simultaneously from the LGN of an anaesthetised marmoset. We have found that noise correlations between parvocellular cells were very low, noise correlations between pairs of magnocellular (M) cells was higher while noise correlations between pairs of koniocellular (K) cells were substantially and significantly higher. This indicates that the connectivity of P cells is ‘designed’ in a way to avoid noise correlations. It is suggested that this lack of substantial noise correlations between P cells subserves their functional role in high acuity vision. Furthermore, it was possible to measure noise correlations between K cells up to 800 μm apart. It is suggested that these noise correlations may result from the topographically diffuse organisation of sub-cortical inputs to the LGN. Chapter 4 investigates a subset of koniocellular neurons in the LGN, which show excitatory responses (generate action potentials) to the presentation of visual stimuli to either eye (binocular cells). While almost all P and M LGN neurons appear to receive excitatory supra-threshold input from one eye only, the binocular cells recorded from K layers exhibit similar receptive field organisation and often respond equally strongly to visual stimuli presented through either eye. The sources and mechanisms underpinning excitatory binocular convergence on substantial proportion of neurons in K layers remain to be established but at the moment the retinal ganglion cells, the superior colliculus and/or parabigeminal nucleus are the most likely suspects. Paucity of excitatory binocular convergence characterizes P cells and M cells of LGN of Old-World primates such as macaque monkeys as well as virtually all LGN cells of eutherian carnivores such as the domestic cat. By contrast, excitatory binocular convergence is one of the predominant characteristics of a majority of LGN neurons in murid rodents. We propose that: 1) a high proportion of binocular cells among koniocellular LGN cells indicates evolutionarily ancient provenance of the primate K pathway and 2) excitatory binocular convergence on K cells combined with the fact that unlike P and M cells, K cells receive direct input from the retino-recipient midbrain nuclei, suggest that this pathway is involved in regulating spatial attention and goal oriented behaviour.