**1. Introduction**

Neuronal migration and precise setting during neurogenesis depend, among others, on reelin (Reln), a 388 kDa glycoprotein secreted by certain neurons within the extracellular matrix [1,2]. The name was given to the protein after the detection of its coding gene, and the acknowledgement that its lack was causative of the mouse *Reeler* mutation [3], which was described, about half a century before, consisting in a form of ataxia [4]. The mutation is autosomic and shows recessive transmission. Consequently, only homozygous recessive *Reeler* mice (*reln*−/−) are totally devoid of Reln and have a definite phenotype. Behaviorally, the latter consists of dystonia, ataxia, and tremor; structurally it primarily affects the design of the cerebral cortex, hippocampus, and cerebellum [5,6]. Contrarily to the mutants, the phenotype of heterozygous *Reeler* mice (*reln*+/−) is normal, but, interestingly, these animals may be translational models of certain human neuropsychiatric disorders [7].

Shortly after the original discovery, it became clear that the mouse gene (*reln*) had a very high homology to that in humans (*RELN*) [8]. Then, a few years later, it was shown that autosomic recessive mutations of the *RELN* gene were linked to a form of lissencephaly with cerebellar hypoplasia (LCH) [9], with associated findings suggested that *RELN* was linked to some neuropsychiatric conditions [10], and *RELN* was demonstrated to be reduced in the cerebellum of autistic patients after Western blotting and immunodetection [11].

Determining a good translational mouse model for a neuropsychiatric condition needs construct, predictive, and face validity [12]. Rigorously, construct validity only relates to transgenic mice, but, in a broader definition, it also comprehends the syndromic models and the spontaneous DNA mutations linked to the phenotype under study. In other words, this factor defines the similarity of the disease between the mouse and the human disorder in terms of the causal gene(s) as e.g., deducted from gene association and linkage analysis. As mentioned above, LCH is a human monogenic condition caused by a mutation in *RELN*. Therefore, the *Reeler* mouse fully meets the criterion of construct validity for the condition. There is also evidence for genetics to be associated with the etiology of several neuropsychiatric conditions, such as autism and schizophrenia, but, as the result of their multidimensional clinical symptoms, causal gene(s), if any, persist to be undiscovered [13]. Nonetheless, there are numerous genes associated with the human autistic pathology after analysis of Mendelian disorders (syndromes), rare mutations, or association studies; see e.g., [14].

Predictive validity, i.e., the similarity of the response to cures in humans and mice is difficult to establish, in the nonexistence of a recognized therapy in humans [14]. Thus, in the context of this discussion, face validity, i.e., the resemblance of the model phenotype to that of the human disorder, is the most important parameter to consider.

Assessment of face validity in neuroscience translational studies requires a careful consideration of their behavioral and structural phenotypes. Broadly speaking, there are contradictory opinions as regarding the repetition in mouse of the human behavioral neuropsychiatric changes. This was, to some extent, predictable, as only a few trials, such as e.g., pre-pulse inhibition (PPI), which records sensory-motor responses, are highly comparable with only minimal modifications in the two species [15]. Notably, the issue has been the subject of several reviews on rodent models of autism, e.g., [16]. The conclusion of these surveys was that, although most of the models that have been used in drug discovery display behaviors with face validity for the human symptoms (i.e., deficits in social communication and restricted interests/repetitive behaviors), many drugs that were found to be useful in ameliorating these autism-related behaviors in mice were ineffective in humans.

Therefore, it becomes imperative to compare the structural alterations of the brains in the two species to substantiate or invalidate the models. We here summarize the state-of-art knowledge on the translational validity of homozygous (*reln*−/−) and heterozygous (*reln*+/−) *Reeler* mice with reference to the most common neuropsychiatric conditions directly or indirectly related to *RELN*. Because of its importance, we will primarily focus onto the brain structural modifications at magnetic resonance imaging (MRI) and histopathology in the two species.
