EXPLORE: a novel deep learning-based analysis method for exploration behaviour in object recognition tests.
Journal
Scientific reports
ISSN: 2045-2322
Titre abrégé: Sci Rep
Pays: England
ID NLM: 101563288
Informations de publication
Date de publication:
14 03 2023
14 03 2023
Historique:
received:
28
07
2022
accepted:
06
03
2023
entrez:
15
3
2023
pubmed:
16
3
2023
medline:
17
3
2023
Statut:
epublish
Résumé
Object recognition tests are widely used in neuroscience to assess memory function in rodents. Despite the experimental simplicity of the task, the interpretation of behavioural features that are counted as object exploration can be complicated. Thus, object exploration is often analysed by manual scoring, which is time-consuming and variable across researchers. Current software using tracking points often lacks precision in capturing complex ethological behaviour. Switching or losing tracking points can bias outcome measures. To overcome these limitations we developed "EXPLORE", a simple, ready-to use and open source pipeline. EXPLORE consists of a convolutional neural network trained in a supervised manner, that extracts features from images and classifies behaviour of rodents near a presented object. EXPLORE achieves human-level accuracy in identifying and scoring exploration behaviour and outperforms commercial software with higher precision, higher versatility and lower time investment, in particular in complex situations. By labeling the respective training data set, users decide by themselves, which types of animal interactions on objects are in- or excluded, ensuring a precise analysis of exploration behaviour. A set of graphical user interfaces (GUIs) provides a beginning-to-end analysis of object recognition tests, accelerating a fast and reproducible data analysis without the need of expertise in programming or deep learning.
Identifiants
pubmed: 36918658
doi: 10.1038/s41598-023-31094-w
pii: 10.1038/s41598-023-31094-w
pmc: PMC10014875
doi:
Types de publication
Journal Article
Research Support, Non-U.S. Gov't
Langues
eng
Sous-ensembles de citation
IM
Pagination
4249Informations de copyright
© 2023. The Author(s).
Références
Behav Brain Res. 1992 Oct 31;51(1):83-92
pubmed: 1482548
Pharmacol Res. 1997 Dec;36(6):463-9
pubmed: 9446713
Nat Protoc. 2013 Dec;8(12):2531-7
pubmed: 24263092
Behav Brain Res. 2001 Dec 14;127(1-2):159-81
pubmed: 11718890
J Neurosci Methods. 1996 Dec 28;70(2):171-5
pubmed: 9007756
J Vis Exp. 2017 Aug 30;(126):
pubmed: 28892027
Nat Neurosci. 2018 Sep;21(9):1281-1289
pubmed: 30127430
Annu Rev Neurosci. 2016 Jul 8;39:217-36
pubmed: 27090952
Elife. 2019 Oct 01;8:
pubmed: 31570119
Neuron. 2019 Oct 9;104(1):11-24
pubmed: 31600508
Nat Methods. 2013 Jan;10(1):64-7
pubmed: 23202433
Nat Protoc. 2006;1(3):1306-11
pubmed: 17406415
Elife. 2021 Nov 30;10:
pubmed: 34846301
Elife. 2021 Sep 02;10:
pubmed: 34473051
Neurosci Biobehav Rev. 2007;31(5):673-704
pubmed: 17368764
J Neurosci Methods. 2008 Mar 15;168(2):422-30
pubmed: 18096240
Neuron. 2014 Oct 1;84(1):18-31
pubmed: 25277452
Eur J Pharmacol. 2007 Mar 8;558(1-3):107-12
pubmed: 17207788
J Pharmacol Toxicol Methods. 2006 Sep-Oct;54(2):99-105
pubmed: 16750402
Neuropsychopharmacology. 2020 Oct;45(11):1942-1952
pubmed: 32711402