Temporal alignment and latent Gaussian process factor inference in population spike trains

We introduce a novel scalable approach to identifying common latent structure in neural population spike-trains, which allows for variability both in the trajectory and in the rate of progression of the underlying computation. Our approach is based on shared latent Gaussian processes (GPs) which are combined linearly, as in the Gaussian Process Factor Analysis (GPFA) algorithm. We extend GPFA to handle unbinned spike-train data by incorporating a continuous time point-process likelihood model, achieving scalability with a sparse variational approximation. Shared variability is separated into terms that express condition dependence, as well as trial-to-trial variation in trajectories. Finally, we introduce a nested GP formulation to capture variability in the rate of evolution along the trajectory. We show that the new method learns to recover latent trajectories in synthetic data, and can accurately identify the trial-to-trial timing of movement-related parameters from motor cortical data without any supervision.

Synaptic transmission modulates while non-synaptic processes govern the transition from pre-ictal to seizure activity in vitro

It is well established that non-synaptic mechanisms can generate electrographic seizures after blockade of synaptic function. We investigated the interaction of intact synaptic activity with non-synaptic mechanisms in the isolated CA1 region of rat hippocampal slices using the “elevated-K+” model of epilepsy. Elevated K+ ictal bursts share waveform features with other models of electrographic seizures, including non-synaptic models where chemical synaptic transmission is suppressed, such as the low-Ca2+model. These features include a prolonged (several seconds) negative field shift associated with neuronal depolarization and superimposed population spikes. When population spikes are disrupted for up to several seconds, intracellular recording demonstrated that the prolonged suppression of population spikes during ictal activity was due to depolarization block of neurons. Elevated-K+ ictal bursts were often preceded by a build-up of “pre-ictal” epileptiform discharges that were characterized as either “slow-transition” (localized and with a gradual increase in population spike amplitude, reminiscent non-synaptic neuronal aggregate formation, presumed mediated by extracellular K+ concentrations ([K+])o accumulation), or “fast-transition” (with a sudden increase in population spike amplitude, presumed mediated by field effects). When ictal activity had a fast-transition it was preceded by fast-transition pre-ictal activity; otherwise population spikes developed gradually at ictal event onset. Addition of bicuculline, a GABAA receptor antagonist, suppressed population spike generation during ictal activity, reduced pre-ictal activity, and increased the frequency of ictal discharges. Nipecotic acid and NNC-711, both of which block GABA re-uptake, increased population spike amplitude during ictal bursts and promoted the generation of preictal activity. By contrast, addition of ionotropic glutamate-receptor antagonists (NBQX, D-APV) had no consistent effect on ictal burst waveform or frequency and did not fully suppress pre-ictal activity. Similarly, CGP 55848, a GABAB receptor antagonist, has no significant effect on pre-ictal activity or burst frequency (although it did increase burst duration slightly). Our results are consistent with the hypothesis that non-synaptic mechanisms underpin the generation of ictal bursts in CA1 and that GABAA synaptic mechanisms can shape event development by delaying event initiation and counteracting depolarization block.