The amygdaloid complex and its associated nuclei play an essential integrative role of sensory, visceral, emotional and behavioral functions. The basolateral nucleus (BLA) is critically important in mediating emotional and adaptive responses to stressors. Virtually no MRI studies report the anatomical characteristics of the BLA, due to the lack of sufficient contrast between the BLA and the surrounding gray matter (GM). Given its relevance we undertook a high-resolution study to obtain BLA volumes in conjunction with imaging approaches to improve contrast to accurately demarcate the BLA within the rodent brain.Eight ex vivo adult rat brains (60d of age) underwent MR imaging using an 11.7 T Bruker Avance. Standard anatomical imaging sequences were obtained, including: 1) T2-weighted imaging (T2WI) sequence with the following parameters: TR/TE=2900/10.2ms (10 echo), 2X2cm FOV, 256² matrix, 4 averages, 0.5mm slice thickness with 25 slices spanning the entire cerebrum; 2) 3D Rapid Acquisition with Relaxation Enhancement (3D RARE) with a 256³ matrix, 2X2 cm FOV, 78 µm slice thickness, 3) An additional DTI scan was acquired on a 9.4 T Bruker with 30 directionally encoded diffusion gradients using a 4-shot EPI sequence with a matrix 128² and then zero-filled to 256² with a TE/TE=12500/36ms, 50 slices at 0.5mm, a 1.92X1.92cm FOV, 4 averages, b value of 0 (5 images) and 3000 s/mm² (30 images, non-collinear directions). This DTI sequence was run prior to the gradient echo DTI for volumetric determinations, and 4) a gradient echo diffusion tensor imaging (DTI) sequence with TR/TE=3000/40ms, 2cm FOV, 2 averages, b=3000 s/mm² and 0.6mm slice thickness with 20 slices covering the brain. . DTI images were analysed using DSI Studio and fractional anisotropy (FA) and primary, secondary and tertiary diffusion eigenvector maps were calculated. Gadolinium (Gd) enhancement was undertaken in an attempt to improve neuroanatomical demarcation, with 3 approaches: 1) 4% PFA fixation only, 2) 4% PFA fixation+1% Gd, and 3) 4% PFA fixation followed by 48hr incubation in 1% Gd with phosphate buffer. Signal (SNR) and contrast to noise ratios (CNR) focusing on the BLA region of the amygdala were assessed. Volumetric BLA and whole brain analysis was performed manually on each coronal slice that contained the BLA using Cheshire image software (Hayden Image/Processing Group) using anatomically defined landmarks. After MR imaging, brains were processed for histological analysis, using cresyl violet staining at 30um sections, using every 10th section.Standard T2WI and high resolution 3D RARE sequences were unable to unequivocally define the anatomical boundaries of the BLA. Similarly, Gadolinium enhancement strategies did not improve identification of the BLA. DTI parametric maps found increased CNR within the BLA region, particularly on the FA and λ1 maps (Fig. 1). Further inspection of each of the individual 30 directionally encoded eigenvector images (0.6mm thickness) and subsequent CNR analysis revealed that the BLA could be optimally observed in 2 specific directions (2, 11) (Fig. 2A). In these orientations, both the right and left BLA had the highest CNR and were visible in multiple slices. For volumetric analysis of the BLA we rescanned the brains using only a single DTI vector profile (11) at 0.6mm/slice, but shifted the slice packet by 0.2mm (3X’s) to obtain an effective slice thickness of 0.2mm/slice. This strategy allowed a 3-fold increase in number of slices for volumetric analysis while still providing sufficient SNR. Using this approach we obtained BLA volumes for the right 1.35±0.12 mm³ and left BLA of 1.39±0.16 mm³ (Fig 2B). MRI derived volumes were validated by comparison to histologically measured BLA volumes which were smaller by 5.7% compared to MRI (Fig. 2C).

MR related studies have demonstrated volumetric alterations in the brain that correlate with neuropsychiatric conditions (i.e. schizophrenia). The amygdala is important in emotional processing, fear, reward, anxiety and cognitive functions with the BLA having critical reciprocal connections to the hippocampus, prefrontal cortex, and other cortical regions. However, there are virtually no volumetric imaging studies evaluating the amygdala and in particular the BLA. We report on a novel strategy where selected directionally encoded DTI images can provide excellent anatomical demarcation from the surrounding GM. We further found that a slice-shifting paradigm with an effective slice thickness of 0.2mm provided accurate volumetric assessments of the BLA in the rodent.

The poor contrast between the BLA and surrounding gray matter has been a barrier in neuroimaging studies. DTI combined with a slice-shifting approach found that we can reliably differentiate the BLA within GM for volumetric studies. Future, clinical and experimental studies could use a similar approach to evaluate the role of the BLA in neuropsychiatric diseases.