We did not detect any significant effects of MS

We did not detect any significant effects of MS on reversal performance in the 4-choice task when mice were tested in adulthood (Fig. 3), even if the daily separation period was extended from 60min to 180min (Fig. 4). This could suggest that early maternal separation has no lasting consequences on decision-making and underlying neural circuits in mice despite extensive evidence of changes in executive function related circuitry in rats (Pascual and Zamora-León, 2007; Chocyk et al., 2010; Monroy et al., 2010; Uchida et al., 2010; Muhammad and Kolb, 2011, Brenhouse et al., 2013; Anier et al., 2014) and degus (Helmeke et al., 2008). This species difference seems unlikely. Yang et al. (2015) recently reported that early maternal stress driven by inadequate supply of bedding produces lasting effects on neuronal morphology in the anterior cingulate acetylcholine chloride in P75 mice and this stress led to impairments in tests of working memory. We also find that in adulthood, mice that had experienced early postnatal maternal separation had higher cumulative ethanol consumption compared to littermates in an intermittent access 20% ethanol single bottle choice “drinking in the dark” paradigm (Fig. 5), suggesting that at least some differences in reward and decision making circuits remain in mice into adulthood. Our ethanol data are consistent with a previous study that found that a 180min maternal separation paradigm increases 10% ethanol intake in a daily three bottle choice paradigm and also increases 6% and 10% ethanol consumption in a daily operant paradigm in adult mice (Cruz et al., 2008). Our ethanol consumption data, acquired using a different paradigm which is able to promote binge-like levels of ethanol drinking in short periods of time, strengthen these previous findings, which together strongly suggest that early maternal separation enhances voluntary ethanol consumption in adulthood. This adds to growing evidence that early adversity may enhance risk for the development of substance use disorders. Our results may also explain why previous studies found no effect of early MS on reversal learning in mice. If, as we observe, the MS effect on flexibility is limited to juvenile development in mice, then effects may have been missed by previous researchers that only tested adults. Task design may also be a factor. Prior studies of MS in mice that found no effect on reversal used the 2-choice attentional set shifting task (Mehta and Schmauss, 2011) while our current study used a 4-choice paradigm. Prior work in rats has shown that the 4-choice reversal task is more difficult to learn than the 2-choice task (Kim and Ragozzino, 2005; Ragozzino et al., 2003; Ragozzino and Rozman, 2007) placing greater cognitive load on the animal to sort out the best of the multiple options. Testing mice with this 4-choice task may have made MS effects easier to detect in mice even when using a supposedly more “stress resistant” (Mehta and Schmauss, 2011) C57Bl/6 strain. Although rats and mice both show similar cognitive differences as a result of early MS, it is notable the age at which MS effects become apparent differs by species. Here we show effects in mice on cognitive flexibility are limited to the juvenile period, while in the literature rats have shown MS effects that emerge after the juvenile period (post P30). For example, working memory deficits assessed using the radial arm maze were observed in maternally separated rats tested at P40, but not at P25 (Brenhouse and Andersen, 2011). Differences in MS rats and controls also emerged after P30 in the Morris water maze (Wang et al., 2015). In this latter study, spatial reversal notably became more flexible in adolescent and adult MS rats compared to controls. This variability in MS effects on cognition suggests there are complex interrelationships between species, age and the specific cognitive domain tested by a task (e.g. spatial vs. non-spatial).