Neural Spike Sorting

Background: Understanding neural functions requires knowledge from analysing electrophysiological data. The process of assigning spikes of a multichannel signal into clusters, called spike sorting, is one of the important problems in such analysis. There have been various automated spike sorting techniques with both advantages and disadvantages regarding accuracy and computational costs. Therefore, developing spike sorting methods that are highly accurate and computationally inexpensive is always a challenge in the biomedical engineering practice.-New method: An automatic unsupervised spike sorting method is proposed in this paper. The method uses features extracted by the locality preserving projection (LPP) algorithm. These features afterwards serve as inputs for the landmark-based spectral clustering (LSC) method. Gap statistics (GS) is employed to evaluate the number of clusters before the LSC can be performed.-Results: The proposed LPP-LSC is a highly accurate and computationally inexpensive spike sorting approach. LPP spike features are very discriminative; thereby boost the performance of clustering methods. Furthermore, the LSC method exhibits its efficiency when integrated with the cluster evaluator GS.-Comparison with existing methods: The proposed method's accuracy is approximately 13% superior to that of the benchmark combination between wavelet transformation and superparamagnetic clustering (WT-SPC). Additionally, LPP-LSC computing time is six times less than that of the WT-SPC.-Conclusions: LPP-LSC obviously demonstrates a win-win spike sorting solution meeting both accuracy and computational cost criteria. LPP and LSC are linear algorithms that help reduce computational burden and thus their combination can be applied into real-time spike analysis.

Knowledge of the activity of single neurons is crucial for understanding neural functions. Therefore the process of attributing every single spike to a particular neuron, called spike sorting, is particularly important in electrophysiological data analysis. This task however is greatly complicated because of numerous factors. Bursts or fast changes in ion channel activation or deactivation can cause a large variability of spike waveforms. Another considerable source of uncertainties results from noise caused by firing of nearby neurons. Movement of electrodes and external electrical noise from the environment also hamper the spike sorting. This paper introduces an integrated approach of diffusion maps (DM), silhouette statistics, and k-means clustering methods for spike sorting. DM is employed to extract spike features that are highly capable of discriminating different spike shapes. The combination of k-means and silhouette statistics provides an automatic unsupervised clustering, which takes features extracted by DM as inputs. Experimental results demonstrate the noticeable superiority of the features extracted by DM compared to those selected by wavelet transformation (WT). Accordingly, the proposed integrated method significantly dominates the popular existing combination of WT and superparamagnetic clustering regarding spike sorting accuracy.

Extracellular data analysis has become a quintessential method for understanding the neurophysiological responses to stimuli. This demands stringent techniques owing to the complicated nature of the recording environment. In this paper, we highlight the challenges in extracellular multi-electrode recording and data analysis as well as the limitations pertaining to some of the currently employed methodologies. To address some of the challenges, we present a unified algorithm in the form of selective sorting. Selective sorting is modelled around hypothesized generative model, which addresses the natural phenomena of spikes triggered by an intricate neuronal population. The algorithm incorporates Cepstrum of Bispectrum, ad hoc clustering algorithms, wavelet transforms, least square and correlation concepts which strategically tailors a sequence to characterize and form distinctive clusters. Additionally, we demonstrate the influence of noise modelled wavelets to sort overlapping spikes. The algorithm is evaluated using both raw and synthesized data sets with different levels of complexity and the performances are tabulated for comparison using widely accepted qualitative and quantitative indicators. Neurophysiological studies are of paramount importance in revealing the underlying behaviours and properties of neurons and eventually providing a good understanding of the human nervous system. The studies have proved to be very important in the development of neuro-prosthetics and Brain Machine Interface (BMI) devices. As an example, intra-neuronal recordings from the primary motor cortex have been investigated to develop neural decoders that can eventually drive artificial prostheses or machines 1. Further, the contribution of these studies in understanding neurological disorders are extremely valued, especially, the use of intracranial electrodes to gather information pertaining to epileptic patients 2. Indeed, MEA's have been employed to understand the influence of gamma-protocadherine, which regulates the endurance of a neural network and the generation of new synapses 3. One of the key aspects of neurophysiological studies involves the tapping of intra-neuronal signals, so as to decipher the neural networks collective behaviours without disrupting their natural functioning. Extracellular recordings are the preferred techniques to aid in neurophysiological studies, and the recordings can be mainly grouped into two categories, that is: in-vivo (invasive) and in-vitro (non-invasive). In-vivo recording techniques use a micro-electrode like probe or a tetrode (probe with four electrodes) to be surgically implanted onto a region of observation, in which intra-neuronal activities are recorded 4. In contrast, in-vitro recording techniques use a micro-electrode array (MEA) with the cell samples cultured in a petri dish 5. Similarly, active cell specimens from animals are collected and placed on the micro-electrodes from which the intra-neuronal activities are recorded 6. Problem Statement Irrespective of the recording techniques used, the intricate nature of the nervous system poses major problems during tapping and processing of intra-neuronal signals. The main attribute of any intra-neuronal activity is the pattern made up of action potential followed by a refractory period, which is referred to as a neuronal spike 7,8. A major problem associated with the processing of any intra-neuronal recording is that each electrode is subjected to more than one neuronal activity at any instance 9. The electrode closer to a neuron renders stronger signals to be picked up by the channel, whilst action potential from neighbouring neurons superimposes upon the stronger