At this very moment, your brain is accomplishing amazing feats. You can see these words and effortlessly understand their meaning. Tomorrow and a week from now you’ll be able to remember parts of what you read. Understanding how the brain encodes, computes and retains information is one of the greatest challenges to neuroscience, and computational modeling is an increasingly important tool in tackling the complexity of the brain and the tasks it accomplishes. Throughout his career, Dr. Terry Sejnowski has contributed hugely to the development of computational neuroscience. His research continues to use innovative computational techniques, combined with experimental data, to elucidate how the brain accomplishes its computation feats. The breath of his lab’s research can begin to be appreciated by looking at their publications from 2019, featuring articles on topics ranging from the role of astrocytic intracellular signaling in long term memory, to identifying feedback projections in early olfactory sensation, to introducing a framework to construct spiking recurrent neural networks that match biological constraints of the cortex and are capable of performing cognitive tasks. An example of Dr. Sejnowski’s lab’s use of computational modeling, anatomical and biological parameters, and experimental data to elucidate neural mechanisms is his lab’s recent publication in The Journal of Neuroscience, “Feedforward Thalamocortical Connectivity Preserves Stimulus Timing Information in Sensory Pathways,” led by Hsi-Ping Wang and Jonathan W. Garcia.
Meaningful sensation and response to the visual world require timing precision and reliability of visual cortex activity. However, it remains incompletely understood how neurons in the primary visual cortex (V1) accomplish this, particularly considering variability of firing in earlier nuclei in the visual pathway, including the Lateral geniculate nucleus (LGN) in the thalamus. (The LGN receives input from retinal ganglion cells (RGC) and relays the visual information to spiny stellate neurons in layer four of V1.) To address this question, Wang et al. (2019) used previously published recordings from LGN neurons in cats (Kara et. al. 2000) as inputs to a Layer-4 spiny stellate cell model (Mainen and Sejnowski, 1998, Figure 1). This allowed them to vary parameters of LGN-V1 connectivity, including the number of LGN inputs, the number of synapses per afferent, and the total number of LGN synapse on the V1 neuron. They were then able to compare the output of the model, its spiking activity, to cat V1 cell spiking data recorded simultaneously to the LGN input by Kara et. al. (2000). Using this model, the authors demonstrated the effects of LGN input within and between trial variability on V1 neuron firing patterns. They found, among other things, that inter-trail variability of LGN firing reduced the reliability and precision of their model’s output, but increasing the number of LGN afferents and intra-trial variability restored reliability and timing precision. Through manipulating the LGN-V1 connectivity parameters of their model, they were able to recapitulate the experimental data and to confirm that the parameters used were consistent with observations from cat V1. This work revealed a novel mechanism by which cortical neurons in the mammalian visual cortex can maintain timing information about visual stimuli. This could provide insight into how thalamocortical inputs preserve stimulus timing information across sensory modalities. To learn more, check out Wang et. al. (2019).
To hear more about Dr. Sejnowski’s approach and recent projects, join us at his talk titled “Lifelong adaptive learning, transfer and savings through gating in the prefrontal cortex” Tuesday April 14, 2020 at 4 pm via Zoom (https://uchealth.zoom.us/j/501283195).
Written by Jennifer Jensen, a 1st year in the Neurosciences Graduate Program at UCSD.
Wang, H.-P., Garcia, J.W., Sabottke, C.F., Spencer, D.J., and Sejnowski, T.J. (2019) Feedforward thalamocortical connectivity preserves stimulus timing information in sensory pathways. J. Neurosci. 39: 7674–7688.
Kara P, Reinagel P, Reid RC (2000) Low response variability in simultaneously recorded retinal, thalamic, and cortical neurons. Neuron 27:635–646.
Mainen ZF, Sejnowski TJ (1995) Reliability of spike timing in neocortical neurons. Science 268:1503–1506.