The prefrontal hemispheres must coordinate dynamically to maintain a unified representation of visual space. Recently, two opposing theories using distinct storage strategies have been proposed: A high-capacity specialized architecture, where each hemisphere governs contralateral behavior, and a fail-safe redundant one, where both hemispheres jointly guide behavior across the visual space. Here, we analyzed simultaneous bilateral prefrontal cortex recordings from three male macaque monkeys performing a visuo-spatial working memory task. Both hemispheres equally predicted behavioral imprecision, decoding errors were weakly correlated between hemispheres, and serial dependence remained local within hemispheres, suggesting a redundant, weakly coupled organization. Attractor network simulations showed that redundancy improved precision when task demands were below memory capacity, while weak interhemispheric coupling increased capacity in more demanding tasks by allowing hemispheric specialization. These predicted patterns were validated in human and monkey data, reconciling previous findings and revealing a versatile interhemispheric architecture that adapts to varying cognitive demands.

An animal’s decision depends not only on incoming sensory evidence but also on its fluctuating internal state. This internal state is a product of cognitive factors, such as fatigue, motivation, and arousal, but it is unclear how these factors influence the neural processes that encode the sensory stimulus and form a decision. We discovered that, over the timescale of tens of minutes during a perceptual decision-making task, animals slowly shifted their likelihood of reporting stimulus changes. They did this unprompted by task conditions. We recorded neural population activity from visual area V4 as well as prefrontal cortex, and found that the activity of both areas slowly drifted together with the behavioral fluctuations. We reasoned that such slow fluctuations in behavior could either be due to slow changes in how the sensory stimulus is processed or due to a process that acts independently of sensory processing. By analyzing the recorded activity in conjunction with models of perceptual decision-making, we found evidence for the slow drift in neural activity acting as an impulsivity signal, overriding sensory evidence to dictate the final decision. Overall, this work uncovers an internal state embedded in the population activity across multiple brain areas, hidden from typical trial-averaged analyses and revealed only when considering the passage of time within each experimental session. Knowledge of this cognitive factor was critical in elucidating how sensory signals and the internal state together contribute to the decision-making process.

Separating neural signals from noise can improve brain-computer interface performance and stability. However, most algorithms for separating neural action potentials from noise are not suitable for use in real time and have shown mixed effects on decoding performance. With the goal of removing noise that impedes online decoding, we sought to automate the intuition of human spike-sorters to operate in real time with an easily tunable parameter governing the stringency with which spike waveforms are classified. We trained an artificial neural network with one hidden layer on neural waveforms that were hand-labeled as either spikes or noise. The network output was a likelihood metric for each waveform it classified, and we tuned the network's stringency by varying the minimum likelihood value for a waveform to be considered a spike. Using the network's labels to exclude noise waveforms, we decoded remembered target location during a memory-guided saccade task from electrode arrays implanted in prefrontal cortex of rhesus macaque monkeys. The network classified waveforms in real time, and its classifications were qualitatively similar to those of a human spike-sorter. Compared with decoding with threshold crossings, in most sessions we improved decoding performance by removing waveforms with low spike likelihood values. Furthermore, decoding with our network's classifications became more beneficial as time since array implantation increased. Our classifier serves as a feasible preprocessing step, with little risk of harm, that could be applied to both off-line neural data analyses and online decoding. NEW & NOTEWORTHY Although there are many spike-sorting methods that isolate well-defined single units, these methods typically involve human intervention and have inconsistent effects on decoding. We used human classified neural waveforms as training data to create an artificial neural network that could be tuned to separate spikes from noise that impaired decoding. We found that this network operated in real time and was suitable for both off-line data processing and online decoding.

A long-standing goal in neuroscience has been to bring together neuronal recordings and neural network modeling to understand brain function. Neuronal recordings can inform the development of network models, and network models can in turn provide predictions for subsequent experiments. Traditionally, neuronal recordings and network models have been related using single-neuron and pairwise spike train statistics. We review here recent studies that have begun to relate neuronal recordings and network models based on the multi-dimensional structure of neuronal population activity, as identified using dimensionality reduction. This approach has been used to study working memory, decision making, motor control, and more. Dimensionality reduction has provided common ground for incisive comparisons and tight interplay between neuronal recordings and network models.

Many studies use population analysis approaches, such as dimensionality reduction, to characterize the activity of large groups of neurons. To date, these methods have treated each neuron equally, without taking into account whether neurons are excitatory or inhibitory. We studied population activity structure as a function of neuron type by applying factor analysis to spontaneous activity from spiking networks with balanced excitation and inhibition. Throughout the study, we characterized population activity structure by measuring its dimensionality and the percentage of overall activity variance that is shared among neurons. First, by sampling only excitatory or only inhibitory neurons, we found that the activity structures of these two populations in balanced networks are measurably different. We also found that the population activity structure is dependent on the ratio of excitatory to inhibitory neurons sampled. Finally we classified neurons from extracellular recordings in the primary visual cortex of anesthetized macaques as putative excitatory or inhibitory using waveform classification, and found similarities with the neuron type-specific population activity structure of a balanced network with excitatory clustering. These results imply that knowledge of neuron type is important, and allows for stronger statistical tests, when interpreting population activity structure.

Adaptive stimulus selection methods in neuroscience have primarily focused on maximizing the firing rate of a single recorded neuron. When recording from a population of neurons, it is usually not possible to find a single stimulus that maximizes the firing rates of all neurons. This motivates optimizing an objective function that takes into account the responses of all recorded neurons together. We propose "Adept," an adaptive stimulus selection method that can optimize population objective functions. In simulations, we first confirmed that population objective functions elicited more diverse stimulus responses than single-neuron objective functions. Then, we tested Adept in a closed-loop electrophysiological experiment in which population activity was recorded from macaque V4, a cortical area known for mid-level visual processing. To predict neural responses, we used the outputs of a deep convolutional neural network model as feature embeddings. Natural images chosen by Adept elicited mean neural responses that were 20% larger than those for randomly-chosen natural images, and also evoked a larger diversity of neural responses. Such adaptive stimulus selection methods can facilitate experiments that involve neurons far from the sensory periphery, for which it is often unclear which stimuli to present.

Recent studies have applied dimensionality reduction methods to understand how the multi-dimensional structure of neural population activity gives rise to brain function. It is unclear, however, how the results obtained from dimensionality reduction generalize to recordings with larger numbers of neurons and trials or how these results relate to the underlying network structure. We address these questions by applying factor analysis to recordings in the visual cortex of non-human primates and to spiking network models that self-generate irregular activity through a balance of excitation and inhibition. We compared the scaling trends of two key outputs of dimensionality reduction—shared dimensionality and percent shared variance—with neuron and trial count. We found that the scaling properties of networks with non-clustered and clustered connectivity differed, and that the in vivo recordings were more consistent with the clustered network. Furthermore, recordings from tens of neurons were sufficient to identify the dominant modes of shared variability that generalize to larger portions of the network. These findings can help guide the interpretation of dimensionality reduction outputs in regimes of limited neuron and trial sampling and help relate these outputs to the underlying network structure.