eprintid: 10199837
rev_number: 9
eprint_status: archive
userid: 699
dir: disk0/10/19/98/37
datestamp: 2024-11-08 08:06:33
lastmod: 2024-11-08 08:06:33
status_changed: 2024-11-08 08:06:33
type: article
metadata_visibility: show
sword_depositor: 699
creators_name: Laranjeira, Simão
creators_name: Guillemot-Legris, Owen
creators_name: Girmahun, Gedion
creators_name: Roberton, Victoria
creators_name: Phillips, James B
creators_name: Shipley, Rebecca J
title: In silico model for automated calculation of functional metrics in animal models of peripheral nerve injury repair
ispublished: pub
divisions: UCL
divisions: B02
divisions: B04
divisions: C08
divisions: D10
divisions: F45
divisions: G10
keywords: Peripheral Nerve Injury; Sciatic mouse model; Machine learning; U-Net; Functional recovery;
Static Sciatic Index; Open Source
note: This version is the author accepted manuscript. For information on re-use, please refer to the publisher’s terms and conditions.
abstract: The rat sciatic nerve model is commonly used to test novel therapies for nerve injury repair. The static sciatic index (SSI) is a useful metric for quantifying functional recovery, and involves comparing an operated paw versus a control paw using a weighted ratio between the toe spread and the internal toe spread. To calculate it, rats are placed in a transparent box, photos are taken from underneath and the toe distances measured manually. This is labour intensive and subject to human error due to the challenge of consistently taking photos, identifying digits and making manual measurements. Although several commercial kits have been developed to address this challenge, they have seen little dissemination due to cost. Here we develop a novel algorithm for automatic measurement of SSI metrics based on video data using casted U-Nets. The algorithm consists of three U-Nets, one to segment the hind paws and two for the two pairs of digits which input into the SSI calculation. A training intersection over union error of 60 % and 80 % was achieved for the back paws and for both digit segmentation U-Nets, respectfully. The algorithm was tested against video data from three separate experiments. Compared to manual measurements, the algorithm provides the same profile of recovery for every experiment but with a tighter standard deviation in the SSI measure. Through the open-source release of this algorithm, we aim to provide an inexpensive tool to more reliably quantify functional recovery metrics to the nerve repair research community.
date: 2024-10
date_type: published
publisher: Elsevier BV
official_url: http://dx.doi.org/10.1016/j.compbiomed.2024.109036
full_text_type: other
language: eng
verified: verified_manual
elements_id: 2309324
doi: 10.1016/j.compbiomed.2024.109036
medium: Print-Electronic
pii: S0010-4825(24)01121-1
lyricists_name: Phillips, James
lyricists_name: Shipley, Rebecca
lyricists_name: Roberton, Victoria
lyricists_name: Laranjeira Gomes, Simao
lyricists_id: JBPHI82
lyricists_id: RSHIP58
lyricists_id: VROBE14
lyricists_id: SJLAR55
actors_name: Laranjeira Gomes, Simao
actors_id: SJLAR55
actors_role: owner
funding_acknowledgements: EP/R004463/1 [Engineering and Physical Sciences Research Council]
full_text_status: restricted
publication: Computers in Biology and Medicine
volume: 181
article_number: 109036
event_location: United States
issn: 0010-4825
citation:        Laranjeira, Simão;    Guillemot-Legris, Owen;    Girmahun, Gedion;    Roberton, Victoria;    Phillips, James B;    Shipley, Rebecca J;      (2024)    In silico model for automated calculation of functional metrics in animal models of peripheral nerve injury repair.                   Computers in Biology and Medicine , 181     , Article 109036.  10.1016/j.compbiomed.2024.109036 <https://doi.org/10.1016/j.compbiomed.2024.109036>.      
 
document_url: https://discovery.ucl.ac.uk/id/eprint/10199837/1/Laranjeira%20Gomes_Manuscript_CBM.pdf