Biology subjects

Majoral, D.

Publications and source records attributed to Majoral, D..

3 recordsLinked to original sources

bioRxiv2024

Kaizen: Decomposing cellular images with VQ-VAE

A fundamental problem in cell and tissue biology is finding cells in microscopy images. Traditionally, this detection has been performed by segmenting the pixel intensities. However, these methods struggle to delineate cells in more densely packed micrographs, where local decisions about boundaries are not trivial. Here, we develop a new methodology to decompose microscopy images into individual cells by making object-level decisions. We formulate the segmentation problem as training a flexible factorized representation of the image. To this end, we introduce Kaizen, an approach inspired by predictive coding in the brain that maintains an internal representation of an image while generating object hypotheses over the external image, and keeping the ones that improve the consistency of internal and external representations. We achieve this by training a Vector Quantised-Variational AutoEncoder (VQ-VAE). During inference, the VQ-VAE is iteratively applied on locations where the internal representation differs from the external image, making new guesses, and keeping only the ones that improve the overall image prediction until the internal representation matches the input. We demonstrate Kaizens merits on two fluorescence microscopy datasets, improving the separation of nuclei and neuronal cells in cell culture images.

bioinformatics↗

bioRxiv2019

Segmenting nuclei in brightfield images with neural networks

Identifying nuclei is a standard first step to analysing cells in microscopy images. The traditional approach relies on signal from a DNA stain, or fluorescent transgene expression localised to the nucleus. However, imaging techniques that do not use fluorescence can also carry useful information. Here, we demonstrate that it is possible to accurately segment nuclei directly from brightfield images using deep learning. We confirmed that three convolutional neural network architectures can be adapted for this task, with U-Net achieving the best overall performance, Mask R-CNN providing an additional benefit of instance segmentation, and DeepCell proving too slow for practical application. We found that accurate segmentation is possible using as few as 16 training images and that models trained on images from similar cell lines can extrapolate well. Acquiring data from multiple focal planes further helps distinguish nuclei in the samples. Overall, our work liberates a fluorescence channel reserved for nuclear staining, thus providing more information from the specimen, and reducing reagents and time required for preparing imaging experiments.

bioinformatics↗

bioRxiv2018

A Model for Time Interval Learning in The Purkinje Cell

Recent experimental findings indicate that Purkinje cells in the cerebellum represent time intervals by mechanisms other than conventional synaptic weights. This finding adds to the theoretical and experimental observations suggesting the presence of intra-cellular mechanisms for adaptation and processing. To account for these experimental results we developed a biophysical model for time interval learning in a Purkinje cell. The numerical model focuses on a classical delay conditioning task (e.g. eyeblink conditioning) and relies on a few computational steps. In particular, the model posits the activation by the parallel fiber input of a local intra-cellular calcium store which can be modulated by intra-cellular pathways. The reciprocal interaction of the calcium signal with several proteins forming negative and positive feedback loops ensures that the timing of inhibition in the Purkinje cell anticipates the interval between parallel and climbing fiber inputs during training. We show that the model is able to learn along the 150-1000 ms interval range. Finally, we discuss how this model would allow the cerebellum to detect and generate specific spatio-temporal patterns, a classical theory for cerebellar function.\n\nAuthor SummaryThe prevailing view in neurosciences considers synaptic weights between neurons the determinant factor for learning and processing information in the nervous system. Theoretical considerations [1, 2] and experiments [3, 4] examined some potential limitations of this classical paradigm, pointing out that adaptation and computation might also have to rely on other mechanisms besides the learning of synaptic weights. Recent experimental findings [5-7] indicate that Purkinje cells in the cerebellum represent time intervals by mechanisms other than conventional synaptic weights. We propose here a biologically plausible model which complements the modification of synaptic weights for learning one time interval in one synapse of one Purkinje cell. In the model a calcium signal in a small domain keeps track of time. Several molecules read and modify this calcium signal to learn a time interval. We discuss how this model would allow the cerebellum to detect and generate specific patterns in space and time, a classical theory for cerebellar function proposed by Braitenberg [8, 9].

neuroscience↗