Ratat1: A Digital Rat Brain Stereotaxic Atlas Derived from High-Resolution MRI Images Scanned in Three Dimensions

摘要

Introduction Numerous magnetic resonance images (MRI)-derived brain atlases are available online for multiple species ( Bakker et al., 2015 ). MRI-derived rat brain atlases have been produced for different purposes using different procedures. The largest number of atlases appear to have been derived for morphometric purposes, e.g., white matter – gray matter segmentation ( Valdes-Hernandez et al., 2011 ), volumetric analysis of major structures ( Papp et al., 2014 ), developmental dynamics ( Calabrese et al., 2013 ). Published rat atlases have drawn their data primarily from albino strains, i.e., Wistar or Sprague Dawley. The present MRI atlas, to our knowledge, is the first available for adult Long-Evans rats, a pigmented strain widely used in behavioral and perceptual research. Some atlases have taken a nomothetic approach, that is to say deriving average templates from groups of animals, e.g. ( Valdes-Hernandez et al., 2011 ) while others have taken an idiographic approach, using data derived from a single exemplar, e.g. ( Papp et al., 2014 ). The nomothetic approach attempts to portray features of a population-average brain, but does so at the risk of statistically blurring small or variable features. The idiographic approach, when combined with high resolution scans, captures small features without statistical blurring, but does so at the risk of including anomalous features. The present atlas was derived from two selected exemplars, thus taking an idiographic approach. In addition, a number of atlases, e.g. ( Papp et al., 2014 ), link or index their images to the widely used reference system established by Paxinos and Watson (1998) (hereafter P&W). This enables users to take advantage of the superior structural detail provided by stained histological sections, and to adopt, if they wish, the widely used P&W coordinate system. An early rat MRI atlas, no longer available online ( Schweinhardt et al., 2003 ), analytically morphed their MRI images to P&W image space using an affine transformation anchored to prominent brain landmarks. From their published examples, line drawings derived from the transformed images were portrayed on a P&W coordinate grid (e.g., Figure 2 ), but the actual MRI images were not (e.g., Figure 3 ) ( Schweinhardt et al., 2003 ). The present atlas did not affine transform the MR images to P&W image space, but rather directly imaged P&W skull landmarks, Bregma and Lambda (reference points visible to a surgeon) and used these to anchor the MR images to the P&W coordinate system via a superimposed grid. The Papp et al. (2014) atlas imaged Bregma and Lambda, but did not project either the reference points or coordinates onto individual brain scans. Animal MRI atlases vary considerably in their method of subject preparation, as well as acquisition parameters affecting image quality, e.g., magnet field strength. Images have been obtained from live subjects ( Valdes-Hernandez et al., 2011 ), ex vivo subjects (as in the present atlas), and preserved tissue ( Papp et al., 2014 ). Each method has advantages and disadvantages. In vivo acquisition obviously captures the brain in its native state, however, movement artifact may affect image quality and anesthesia tolerance limits total acquisition time (which also limits image resolution). Images obtained from preserved tissue eliminate the issues of movement artifact and acquisition time but bring with it structural changes from preservative perfusion. The ex vivo method, used for the present atlas, maintains in situ structure, but acquisition time is constrained by post-mortem tissue degradation (approximately 10 h in the present environment). Higher magnet field strength brings with it improved spatial resolution, but also imaging artifacts that accrue in proportion to field strength (discussed below). Objective To derive a high resolution digital brain atlas of in situ MR images from adult Long Evans rats suitable for use in stereotaxic surgery applications. Using an ultra-high resolution scanner, brain images were acquired ex vivo in a near natural state. Those images, particularly when combined with images derived from conventional histological sections, can be used to determine the coordinates of stereotaxic surgical targets with improved accuracy. Additional objectives were to depict brain images in three viewing planes indexed to the skull surface as well as to the skull landmarks Bregma and Lambda. A further objective was to present the atlas in a simple widely used file format (.pdf) not requiring specialized software for viewing, copying, or unpacking. Subjects Nine adult Long Evans (Harlan, Indianapolis, IN, USA) male rats, wt. 350–550 g were imaged. The image sets of two animals were judged to be of superior quality and were used as exemplars to compose the atlas. Their images were free of anatomical anomalies and imaging distortions. These animals were 135–136 days of age when imaged, and weighed 506 and 510 g. The experimental