SimpSyn

Short overview of what this module does and links to usage.

• Install & Usage: See the module's README and scripts in GitHub to get more up-to-date information.
• Design Choices: See Design Choices for the "why" behind the "what".

For a primer to "What synapse detection is?", check here.

What is SimpSyn?

SimpSyn is our ultra-light weight model for synapse detection. The name itself stands for Simple Synapse Detection.

With SimpSyn we rethink the synapse detection problem as a segmentation task, where both pre and post sites are segmented as masks and are post-processed into partners via nearest neighbour assignments. This deviates from Synful's joint learning of a mask segmentation and regression tasks, thereby being computationally efficient.

SimpSyn uses a 3D Residual U-Net for synapse detection that predicts two output channels: one corresponding to pre-synaptic regions and the other to post-synaptic regions. To achieve this, a spherical 3D region is generated and centered at the coordinates of the pre-synaptic and post-synaptic points. These output masks are subsequently processed using connected component labelling to isolate individual synaptic structures. To establish correspondence between pre-synaptic and post-synaptic sites, each post-synaptic component is paired with its nearest pre-synaptic counterpart based on the nearest neighbour criterion.

Getting Started

SIMPSYN is built within Biapy. Please follow the installation instructions here.

Kindly note SimpSyn is currently being tested on various other datasets as our paper is under review.

Complete documentation will be released soon.

Citing SIMPSYN

Please cite our preprint on Towards Generalized Synapse Detection Across Invertebrate Species.