A novel ex vivo MR imaging template of the Japanese quail to study stress-selected lines.
David André Barrière1, Raïssa Yebga Hot1, Marine Siwiaszczyk2, Justine Beaujoin1, Ivy Uszynski1, Scott Love2, Baptiste Mulot3, Ludovic Calandreau2, Élodie Chaillou2, and Cyril Poupon1

1Neurospin, CEA, Gif-sur-Yvette, France, 2PRC, INRA CNRS IFCE, Nouzilly, France, 3Zooparc Beauval & Beauval Nature, Saint-Aignan, France


In neuroscience, birds are becoming interesting animal models to study learning and memory but also response to stress. Nevertheless, bird’s brain organization and physiology suffers from a lack of neuroimaging tools to perform non-invasive and longitudinal investigations. In this study, we proposed a novel brain template of the Japanese quail (Coturnix Japonica), built from twenty animals dedicated to voxel-based morphometry approach. Using these tools we investigate differences in grey matter concentration (GMC) between two divergent lines of quails selected from their response to fear. Our results report structural differences between the both quail lineages within cognitives, motivational and motor systems.


Japanese quails (Coturnix Japonica) genetically selected for long (LTI) rather than short (STI) tonic immobility reactions show a more pronounced fear-responses in LTI animals in frightening situations1. Nevertheless, no clear evidence reporting disturbance of the hypothalamo-pituitary-adrenal axis has been in reported this line suggesting that higher central modifications support the phenotype1. Hence, morphological investigations of quail’s brains using anatomical MRI gained in interest to map structural difference between both lineage. Nevertheless, no specific neuroinformatic tools dedicated to Japanese quails have been published2–4. Using ultra high field MRI (11.7 T) and a unified segmentation approach5–7, we generated an brain template of the Japanese quail and its corresponding tissue probability maps to explore brain modifications.

Materials and methods

Japanese quail cohort - Twenty male Japanese quails, including 10 LTI and 10 STI subjects, were euthanized, perfused using 4% paraformaldehyde solution, heads were collected and conserved in PFA. Three days before scanning heads were rehydrated and placed in Fluor Inert.

MRI protocols - The subjects were scanned with a 11.7T MRI system (Bruker) equipped using a 40 mm Bruker 1H transmit-receive volume coil. The imaging protocol included a T2-weighted spin echo (SE) sequence ( 150x150x150 μm, TE=16ms, TR=9000ms, 8 averages.

Template and priors building - Brains were realigned to the AC/CP axis and realigned to the first animal. First, each animal was segmented in grey matter (GM), white matter (WM) and cerebrospinal fluid (CSF) probability maps using the FAST function of FSL. GM and WM priors were used to generate the first template using a DARTEL approach5. In a second step, these preliminary probability maps were used to segment and normalized the native, coregistered images. A second run of DARTEL was performed to generate final probabilistic maps but also Jacobian images containing the deformation parameters which were used to warp each brains6. Finally, warped images were averaged to create the final ex vivo template.

Atlas building - Brain regions were identified and delineated manually thanks to the brain template and priors’ maps generated previously. Each prior was binarized before segmenting 46 cortical/subcortical brain regions within the GM mask using the ITK toolkit. The WM and CSF masks were finally labelled separately and added to the atlas.

VBM and statistical analysis - Previously pre-processed normalized T2 data were segmented from the previously generated priors’ images non linearly normalized to the brain template using DARTEL. GM images were modulated by the determinant of the Jacobian then spatially smoothed with a 3 mm isotropic Gaussian kernel to create grey matter concentration maps (GMC). Regional changes between GMC maps of LTI versus STI were revealed by a two sample Student comparison test (p < 0.05 (t(18) = 1.77).

Results and discussion

Here, we built a high-resolution (150 μm isometric voxel) anatomical template of the Japanese quail brain (Figure 1A). Accompanying our template, we propose tissue probability maps (TPMs) of GM, WM and CSF (Figure 1B, C, D). Comparison between the GMC maps of GM between LTI and STI lineage reveal significant variations of GMC with local decreasing of GM especially within cortical regions and few increasing of GM within the diencephalon (Figure 2). To identify the brain regions supporting the GMC modifications we established a brain atlas of the Japanese quails accordingly to 2–4. Forty-six ROIs delineating the arcopallium, hyperpallium, mesopallium and nidopallium as well as the hippocampus and the striatum were defined within the GM mask. Further, brain regions such as the olfactory bulb, midbrain, diencephalon, cerebellum and the pons and nuclei such as the preoptic medial area, hypothalamus, or the ventral tegmental area (Figure 3) were also delineated. Using this brain atlas, we identified 22 significant clusters localized cortically in the nidopallium, arcopallium, mesopallium, hyperpallium and subcortically within the diencephalon, striatum and cerebellum (Table 1). Preliminar results show that the WM volume is significantly lower in the STI quails resulting to a significant increasing of the GM/WM ratio in the STI quails (Figure 4).


The Japanese quail template and its associated probabilistic maps created in this work offer a novel tool allowing to investigate the avian brain physiology and differences between STI and LTI quail lines. Here, we demonstrated that discrete modifications of the GMC are present in LTI within regions involved in the cognitive process but also in regions involved in motivational and motor systems. Last, our study showed that the WM volume is increased in animals stemming from the STI line, which has still to be confirmed using both diffusion MRI and histological studies.


This study received funding from the Centre-Val de Loire regional council (Neuro2Co, France), Beauval Nature and the RHU TRT_cSVD.


1. Jones, R. B., Mills, A. D., Faure, J. M. & Williams, J. B. Restraint, fear, and distress in Japanese quail genetically selected for long or short tonic immobility reactions. Physiol. Behav. 56, 529–534 (1994).

2. Vellema, M., Verschueren, J., Van Meir, V. & Van der Linden, A. A customizable 3-dimensional digital atlas of the canary brain in multiple modalities. Neuroimage 57, 352–361 (2011).

3. Güntürkün, O., Verhoye, M., De Groof, G. & Van der Linden, A. A 3-dimensional digital atlas of the ascending sensory and the descending motor systems in the pigeon brain. Brain Struct Funct 218, 269–281 (2013).

4. De Groof, G. et al. A three-dimensional digital atlas of the starling brain. Brain Struct Funct 221, 1899–1909 (2016).

5. Ashburner, J. A fast diffeomorphic image registration algorithm. Neuroimage 38, 95–113 (2007).

6. Ashburner, J. & Friston, K. J. Voxel-based morphometry--the methods. Neuroimage 11, 805–821 (2000).

7. Ashburner, J. & Friston, K. J. Unified segmentation. Neuroimage 26, 839–851 (2005).


Axial views of the ex vivo Japanese Quail template (A) and its associated tissue probability maps Gray Matter (B), White Matter (C) and CSF (D). All the images have been linearly registered to the AC-PC axis for conventional orientation.

Brain slices showing the local regional grey matter concentration (GMC) modifications between Long-Term Immobilization (LTI) and Short-Term Immobilization (STI) lines. SPM two sample Student comparison test (voxel level threshold p < 0.05, t(18) = 1.77, cluster threshold = 100 voxels, LTI, n = 10, STI n = 10)

Coronal (A) and sagittal (B) slices of the ex vivo Japanese quail atlas and 3D representations of the atlas (C).

Cluster identification and summary statistics of VBM analysis. SPM two sample Student comparison test (voxel level threshold p < 0.05, t(18) = 1.77, cluster threshold = 100 voxels, LTI, n = 10, STI n = 10).

Comparison of total (A) grey matter (B) and white matter (C) volumes and GM/WM ratio between the STI and LTI lines. STI (n = 10) and LTI (n = 10) data were expressed as mean ± SEM; and compared using a Student t-test * p < 0.05.

Proc. Intl. Soc. Mag. Reson. Med. 27 (2019)