Stephanie E. Palmer

The Neural Basis for Combinatorial Coding in a Cortical Population Response

We have used a combination of theory and experiment to assess how information is represented in a realistic cortical population response, examining how motion direction and timing is encoded in groups of neurons in cortical area MT. Combining data from several single-unit experiments, we constructed model population responses in small time windows and represented the response in each window as a binary vector of 1s or 0s signifying spikes or no spikes from each cell. We found that patterns of spikes and silence across a population of nominally redundant neurons can carry up to twice as much information about visual motion than does population spike count, even when the neurons respond independently to their sensory inputs. This extra information arises by virtue of the broad diversity of firing rate dynamics found in even very similarly tuned groups of MT neurons. Additionally, specific patterns of spiking and silence can carry more information than the sum of their parts (synergy), opening up the possibility for combinatorial coding in cortex. These results also held for populations in which we imposed levels of nonindependence (correlation) comparable to those found in cortical recordings. Our findings suggest that combinatorial codes are advantageous for representing stimulus information on short time scales, even when neurons have no complicated, stimulus-dependent correlation structure.

Predictive information in a sensory population.

Significance Prediction is an essential part of life. However, are we really “good” at making predictions? More specifically, are pieces of our brain close to being optimal predictors? To assess the efficiency of prediction, we need to measure the information that neurons carry about the future of our sensory experiences. We show how to do this, at least in simplified contexts, and find that groups of neurons in the retina indeed are close to maximally efficient at separating predictive information from the nonpredictive background. Efficient coding of predictive information is a principle that can be applied at every stage of neural computation. Guiding behavior requires the brain to make predictions about the future values of sensory inputs. Here, we show that efficient predictive computation starts at the earliest stages of the visual system. We compute how much information groups of retinal ganglion cells carry about the future state of their visual inputs and show that nearly every cell in the retina participates in a group of cells for which this predictive information is close to the physical limit set by the statistical structure of the inputs themselves. Groups of cells in the retina carry information about the future state of their own activity, and we show that this information can be compressed further and encoded by downstream predictor neurons that exhibit feature selectivity that would support predictive computations. Efficient representation of predictive information is a candidate principle that can be applied at each stage of neural computation.

Beta cells cannot directly prime diabetogenic CD8 T cells in nonobese diabetic mice.

Type 1 diabetes (T1D) is caused by the destruction of insulin-producing islet β cells. CD8 T cells are prevalent in the islets of T1D patients and are the major effectors of β cell destruction in nonobese diabetic (NOD) mice. In addition to their critical involvement in the late stages of diabetes, CD8 T cells are implicated in the initiation of disease. NOD mice, in which the β2-microglobulin gene has been inactivated by gene targeting (NOD.β2M−/−), have a deficiency in CD8 T cells and do not develop insulitis, which suggests that CD8 T cells are required for the initiation of T1D. However, neither in humans nor in NOD mice have the immunological requirements for diabetogenic CD8 T cells been precisely defined. In particular, it is not known in which cell type MHC class I expression is required for recruitment and activation of CD8 T cells. Here we have generated transgenic NOD mice, which lack MHC class I on mature professional antigen-presenting cells (pAPCs). These “class I APC-bald” mice developed periislet insulitis but not invasive intraislet insulitis, and they never became diabetic. Recruitment to the islet milieu does not therefore require cognate interaction between CD8 T cells and MHC class I on mature pAPCs. Conversely, such an interaction is critically essential to allow the crucial shift from periislet insulitis to invasive insulitis. Importantly, our findings demonstrate unequivocally that CD8 T cells cannot be primed to become diabetogenic by islet β cells alone.

Quantum disorder in the two-dimensional pyrochlore Heisenberg antiferromagnet

We present the results of an exact diagonalization study of the spin-

Normal mode oscillation in the presence of inhomogeneous broadening.

1/2

Tracing the origin and evolution of supergene mimicry in butterflies

Heisenberg antiferromagnet on a two-dimensional version of the pyrochlore lattice, also known as the square lattice with crossings or the checkerboard lattice. Examining the low-energy spectra for systems of up to 24 spins, we find that all clusters studied have nondegenerate ground states with total spin zero, and big energy gaps to states with higher total spin. We also find a large number of nonmagnetic excitations at energies within this spin gap. Spin-spin and spin-Peierls correlation functions appear to be short ranged, and we suggest that the ground state is magnetically disordered.

Decoding thalamic afferent input using microcircuit spiking activity

We investigate effects of inhomogeneous broadening of excitons on normal mode oscillation in semiconductor microcavities using a coupled oscillator model. We show that inhomogeneous broadening can drastically alter the coherent oscillatory energy exchange process even in regimes where normal mode splitting remains nearly unchanged. The depth, frequency, and phase of normal mode oscillations of excitons at a given energy within the inhomogeneous distribution depend strongly on the energy separation between the exciton and the normal mode resonance. In addition, for an inhomogeneous broadened system, pronounced oscillations in the intensity of the optical field or the total induced optical polarization no longer imply a similar oscillatory coherent energy exchange between excitons and cavity photons.

Learning to make external sensory stimulus predictions using internal correlations in populations of neurons

Supergene mimicry is a striking phenomenon but we know little about the evolution of this trait in any species. Here, by studying genomes of butterflies from a recent radiation in which supergene mimicry has been isolated to the gene doublesex, we show that sexually dimorphic mimicry and female-limited polymorphism are evolutionarily related as a result of ancient balancing selection combined with independent origins of similar morphs in different lineages and secondary loss of polymorphism in other lineages. Evolutionary loss of polymorphism appears to have resulted from an interaction between natural selection and genetic drift. Furthermore, molecular evolution of the supergene is dominated not by adaptive protein evolution or balancing selection, but by extensive hitchhiking of linked variants on the mimetic dsx haplotype that occurred at the origin of mimicry. Our results suggest that chance events have played important and possibly opposing roles throughout the history of this classic example of adaptation.Wing pattern mimicry in the butterfly Papilio polytes is controlled by a single Mendelian locus, the mimicry supergene doublesex. Here, Zhang and colleagues reconstruct the complex evolutionary history of the doublesex supergene and mimicry in the Papilio polytes species group.

Transformation of Stimulus Correlations by the Retina

A behavioral response appropriate to a sensory stimulus depends on the collective activity of thousands of interconnected neurons. The majority of cortical connections arise from neighboring neurons, and thus understanding the cortical code requires characterizing information representation at the scale of the cortical microcircuit. Using two-photon calcium imaging, we densely sampled the thalamically evoked response of hundreds of neurons spanning multiple layers and columns in thalamocortical slices of mouse somatosensory cortex. We then used a biologically plausible decoder to characterize the representation of two distinct thalamic inputs, at the level of the microcircuit, to reveal those aspects of the activity pattern that are likely relevant to downstream neurons. Our data suggest a sparse code, distributed across lamina, in which a small population of cells carries stimulus-relevant information. Furthermore, we find that, within this subset of neurons, decoder performance improves when noise correlations are taken into account.

State dependence of stimulus-induced variability tuning in macaque MT

Significance To produce appropriate behavioral responses, such as catching fast-moving prey, the visual system copes with significant sensory processing delays. Spiking activity in the retina captures some of the most predictive aspects of the visual information, but this information must be accessible to downstream circuits. We tested how efficiently predictive information could be read out in downstream neurons and how difficult it is to learn to read out this information, using biologically plausible rules applied only to local inputs. Very simple learning rules could find near-optimal readouts of predictive information without any external instructive signal. This suggests that bottom-up prediction may play an important role in sensory processing. To compensate for sensory processing delays, the visual system must make predictions to ensure timely and appropriate behaviors. Recent work has found predictive information about the stimulus in neural populations early in vision processing, starting in the retina. However, to utilize this information, cells downstream must be able to read out the predictive information from the spiking activity of retinal ganglion cells. Here we investigate whether a downstream cell could learn efficient encoding of predictive information in its inputs from the correlations in the inputs themselves, in the absence of other instructive signals. We simulate learning driven by spiking activity recorded in salamander retina. We model a downstream cell as a binary neuron receiving a small group of weighted inputs and quantify the predictive information between activity in the binary neuron and future input. Input weights change according to spike timing–dependent learning rules during a training period. We characterize the readouts learned under spike timing–dependent synaptic update rules, finding that although the fixed points of learning dynamics are not associated with absolute optimal readouts they convey nearly all of the information conveyed by the optimal readout. Moreover, we find that learned perceptrons transmit position and velocity information of a moving-bar stimulus nearly as efficiently as optimal perceptrons. We conclude that predictive information is, in principle, readable from the perspective of downstream neurons in the absence of other inputs. This suggests an important role for feedforward prediction in sensory encoding.