On comparison of manifold learning techniques for dendritic spine classification

Abstract

Dendritic spines are one of the key functional components of neurons. Their morphological changes are correlated with neuronal activity. Neuroscientists study spine shape variations to understand their relation with neuronal activity. Currently this analysis performed manually, the availability of reliable automated tools would assist neuroscientists and accelerate this research. Previously, morphological features based spine analysis has been performed and reported in the literature. In this paper, we explore the idea of using and comparing manifold learning techniques for classifying spine shapes. We start with automatically segmented data and construct our feature vector by stacking and concatenating the columns of images. Further, we apply unsupervised manifold learning algorithms and compare their performance in the context of dendritic spine classification. We achieved 85.95% accuracy on a dataset of 242 automatically segmented mushroom and stubby spines. We also observed that ISOMAP implicitly computes prominent features suitable for classification purposes.

Dendritic Spine Shape Analysis: A Clustering Perspective

“…8: Average image for each cluster generated using the DNSM features. using information gain [40] and conclude that neck length is the most dominant feature for data used in this study, which confirms analysis performed in some of our previous studies [25,30].…”

“…Shi et al [7] developed a semi-supervised learning approach for spine classification based on morphological features, and used human experts for validation of their results. A recent study on spine analysis applied ISOMAP [29] to study the importance of different morphological parameters and found neck length and head diameter to be the most prominent features for mushroom and stubby spines [30]. Ghani et al [31] exploited the parametric nature of the DNSM approach and used its parameters for spine classification; they also used labels assigned by a human expert for performance evaluation.…”