Christophe Pouzat

Using noise signature to optimize spike-sorting and to assess neuronal classification quality

We have developed a simple and expandable procedure for classification and validation of extracellular data based on a probabilistic model of data generation. This approach relies on an empirical characterization of the recording noise. We first use this noise characterization to optimize the clustering of recorded events into putative neurons. As a second step, we use the noise model again to assess the quality of each cluster by comparing the within-cluster variability to that of the noise. This second step can be performed independently of the clustering algorithm used, and it provides the user with quantitative as well as visual tests of the quality of the classification.

Developmental regulation of basket/stellate cell->Purkinje cell synapses in the cerebellum

We used paired recordings to study the development of synaptic transmission between inhibitory interneurons of the molecular layer and Purkinje cells in the cerebellar cortex of the rat. The electrophysiological data were combined with a morphological study of the recorded cells using biocytin or Lucifer yellow staining. Thirty-one interneuron–Purkinje cell pairs were obtained, and 11 of them were recovered morphologically. The age of the rats ranged from 11 to 31 d after birth. During this period synaptic maturation resulted in an 11-fold decrease in the average current evoked in a Purkinje cell by a spike in a presynaptic interneuron. Unitary IPSCs in younger animals exhibited paired-pulse depression, whereas paired-pulse facilitation was found in more mature animals. These data suggest that reduction in transmitter release probability contributed to the developmental decrease of unitary IPSCs. However, additional mechanisms at both presynaptic and postsynaptic loci should also be considered. The decrease of the average synaptic current evoked in a Purkinje cell by an action potential in a single interneuron suggests that as development proceeds interneuron activities must be coordinated to inhibit efficiently Purkinje cells.

Towards reliable spike-train recordings from thousands of neurons with multielectrodes.

The new generation of silicon-based multielectrodes comprising hundreds or more electrode contacts offers unprecedented possibilities for simultaneous recordings of spike trains from thousands of neurons. Such data will not only be invaluable for finding out how neural networks in the brain work, but will likely be important also for neural prosthesis applications. This opportunity can only be realized if efficient, accurate and validated methods for automatic spike sorting are provided. In this review we describe some of the challenges that must be met to achieve this goal, and in particular argue for the critical need of realistic model data to be used as ground truth in the validation of spike-sorting algorithms.

Action potential‐evoked Ca2+ signals and calcium channels in axons of developing rat cerebellar interneurones

1 Axonal [Ca2+] transients evoked by action potential (AP) propagation were studied by monitoring the fluorescence of the high‐affinity calcium‐sensitive dye Oregon Green 488 BAPTA‐1, introduced through whole‐cell recording pipettes in the molecular layer of interneurones from cerebellar slices of young rats. 2 The spatiotemporal profile of Ca2+‐dependent fluorescence changes was analysed in well‐focused axonal stretches a few tens of micrometres long. AP‐evoked Ca2+ signals were heterogeneously distributed along axons, with the largest and fastest responses appearing in hot spots on average ∼5 μm apart. 3 The spatial distribution of fluorescence responses was independent of the position of the focal plane, uncorrelated to basal dye fluorescence, and independent of dye concentration. Recordings using the low‐affinity dye mag‐fura‐2 and a Cs+‐based intracellular solution revealed a similar pattern of hot spots in response to depolarisation, ruling out measurement artefacts or possible effects of inhomogeneous dye distribution in the generation of hot spots. 4 Fluorescence responses to a short train of APs in hot spots decreased by 41–76 % after bath perfusion of ω‐conotoxin MVIIC (5–6 μM), and by 17–65 % after application of ω‐agatoxin IVA (500 nM). ω‐Conotoxin GVIA (1 μM) had a variable, small effect (0–31 % inhibition), and nimodipine (5 μM) had none. Somatically recorded voltage‐gated currents during depolarising pulses were unaffected in all cases. These data indicate that P/Q‐type Ca2+ channels, and to a lesser extent N‐type channels, are responsible for a large fraction of the [Ca2+] rise in axonalhot spots. 5 [Ca2+] responses never failed during low‐frequency (≤ 0.5 Hz) stimulation, indicating reliable AP propagation to the imaged sites. Axonal branching points coincided with a hot spot in ∼50 % of the cases. 6 The spacing of presynaptic varicosities, as determined by a morphological analysis of Neurobiotin‐filled axons, was ∼10 times larger than the one measured for hot spots. The latter is comparable to the spacing reported for varicosities in mature animals. 7 We discuss the nature of hot spots, considering as the most parsimonious explanation that they represent functional clusters of voltage‐dependent Ca2+ channels, and possibly other [Ca2+] sources, marking the position of developing presynaptic terminals before the formation of en passant varicosities.

Autaptic inhibitory currents recorded from interneurones in rat cerebellar slices

1 While the presence of autapses in the brain is indicated by a large body of morphological evidence, the functional role of these structures has remained unclear. To probe for autaptic currents, we have recorded current responses following short somatic depolarizing pulses in Cl−‐loaded interneurones (stellate and basket cells) from rat cerebellar slices (animals aged 27‐39 days). 2 In ≈20 % of the recordings, fluctuating inward current transients were obtained with a latency of 1.15‐2.45 ms (measured from the peak of the depolarization‐induced Na+ current; n= 10). 3 These transients were blocked by bicuculline and were sensitive to the extracellular Ca2+ concentration. 4 Assuming low release probability, as suggested by the high failure rate (0.65‐0.92, n= 10), quantal sizes ranging from 21 to 178 pA (‐70 mV; n= 10) were calculated from a variance analysis of autaptic current amplitudes. 5 We conclude that ≈20 % of interneurones have a functional autapse. Autaptic currents may inhibit firing of interneurones during high frequency discharges.

An Algebraic Method for Eye Blink Artifacts Detection in Single Channel EEG Recordings

Single channel EEG systems are very useful in EEG based applications where real time processing, low computational complexity and low cumbersomeness are critical constrains. These include brain-computer interface and biofeedback devices and also some clinical applications such as EEG recording on babies or Alzheimer’s disease recognition. In this paper we address the problem of eye blink artifacts detection in such systems. We study an algebraic approach based on numerical differentiation, which is recently introduced from operational calculus. The occurrence of an artifact is modeled as an irregularity which appears explicitly in the time (generalized) derivative of the EEG signal as a delay. Manipulating such delay is easy with the operational calculus and it leads to a simple joint detection and localization algorithm. While the algorithm is devised based on continuous-time arguments, the final implementation step is fully realized in a discrete-time context, using very classical discrete-time FIR filters. The proposed approach is compared with three other approaches: (1) the very basic threshold approach, (2) the approach that combines the use of median filter, matched filter and nonlinear energy operator (NEO) and (3) the wavelet based approach. Comparison is done on: (a) the artificially created signal where the eye activity is synthesized from real EEG recordings and (b) the real single channel EEG recordings from 32 different brain locations. Results are presented with Receiver Operating Characteristics curves. The results show that the proposed approach compares to the other approaches better or as good as, while having lower computational complexity with simple real time implementation. Comparison of the results on artificially created and real signal leads to conclusions that with detection techniques based on derivative estimation we are able to detect not only eye blink artifacts, but also any spike shaped artifact, even if it is very low in amplitude.

Axonal Speeding: Shaping Synaptic Potentials in Small Neurons by the Axonal Membrane Compartment

The role of the axonal membrane compartment in synaptic integration is usually neglected. We show here that in interneurons of the cerebellar molecular layer, where dendrites are so short that the somatodendritic domain can be considered isopotential, the axonal membrane contributes a significant part of the cell input capacitance. We examine the impact of axonal membrane on synaptic integration by cutting the axon with two-photon illumination. We find that the axonal compartment acts as a sink for signals generated at fast conductance synapses, thus increasing the initial decay rate of corresponding synaptic potentials over the value predicted from the resistance-capacitance (RC) product of the cell membrane; signals generated at slower synapses are much less affected. This mechanism sharpens the spike firing precision of fast glutamatergic inputs without resorting to multisynaptic pathways.

Toward standard practices for sharing computer code and programs in neuroscience

Computational techniques are central in many areas of neuroscience and are relatively easy to share. This paper describes why computer programs underlying scientific publications should be shared and lists simple steps for sharing. Together with ongoing efforts in data sharing, this should aid reproducibility of research.

Quantitative estimation of calcium dynamics from ratiometric measurements: a direct, nonratioing method.

Measuring variations of intracellular free calcium concentration through the changes in fluorescence of a calcium-sensitive dye is a ubiquitous technique in neuroscience. Despite its popularity, confidence intervals (CIs) on the estimated parameters of calcium dynamics models are seldom given. To address this issue, we have developed a two-stage model for ratiometric measurements obtained with a charge-coupled device (CCD) camera. Its first element embeds a parametric calcium dynamics model into a fluorescence intensity model and its second element probabilistically describes the fluorescence measurements by a CCD camera. Using Monte Carlo simulations, we first show that the classical ratiometric transformation gives reliable CIs for time constants only and not baseline calcium concentration nor influx. We then introduce a direct method, which consists of fitting directly and simultaneously the fluorescence transients at both wavelengths, without any data ratioing. This approach uses a probabilistic description of the camera, leading to the construction of meaningful CIs for the calcium parameters. Moreover, using approaches inspired by constrained linear regression, we can take into account the finite precision on calibrated parameters (such as the dye dissociation constant in the cell). These key features are illustrated on simulated data using Monte Carlo simulations. Moreover, we illustrate the strength of the direct method on experimental recordings from insect olfactory interneurons. In particular, we show how to handle a time-dependent buffer concentration, thereby considerably improving our goodness of fit. The direct method was implemented in the open-source software R and is freely distributed in the CalciOMatic package.

SIMONE: A Realistic Neural Network Simulator to Reproduce MEA-Based Recordings

Contemporary multielectrode arrays (MEAs) used to record extracellular activity from neural tissues can deliver data at rates on the order of 100 Mbps. Such rates require efficient data compression and/or preprocessing algorithms implemented on an application specific integrated circuit (ASIC) close to the MEA. We present SIMONE (Statistical sIMulation Of Neuronal networks Engine), a versatile simulation tool whose parameters can be either fixed or defined by a probability distribution. We validated our tool by simulating data recorded from the first olfactory relay of an insect. Different key aspects make this tool suitable for testing the robustness and accuracy of neural signal processing algorithms (such as the detection, alignment, and classification of spikes). For instance, most of the parameters can be defined by a probabilistic distribution, then tens of simulations may be obtained from the same scenario. This is especially useful when validating the robustness of the processing algorithm. Moreover, the number of active cells and the exact firing activity of each one of them is perfectly known, which provides an easy way to test accuracy.