“…According to Nylander and Roman (2013), the consequence of the variability in protocols is that there is no single 'rodent maternal separation model' but several separation models with different effects. Although separation models are not safe from criticism (Molet et al, 2014), a large body of ELS evidence indicates that postnatal manipulations are important models for investigating the early-life impact on depressive and anxiety-like behaviours (Andersen, 2015;Millstein and Holmes, 2007;Romeo et al, 2003;Schmidt et al, 2011;Vetulani, 2013), drug-seeking behaviours (Cruz et al, 2008;Garcia-Gutierrez et al, 2015;Huot et al, 2001;Kawakami et al, 2007;Nylander and Roman, 2013) and neuroendocrine stressinduced responses among rodents (Anisman et al, 1998;Kawakami et al, 2013;Loizzo et al, 2010;Parfitt et al, 2004;Parfitt et al, 2007). The evidence has provided reliable implications for face, construct and predictive validity for MS rodent studies (Plotsky et al, 2005;Schmidt et al, 2011).…”