Maternal Immune Activation and its Multifaceted Effects on Learning and Memory in Rodent Offspring: A Systematic Review

This systematic review explored the impact of maternal immune activation (MIA) on learning and memory behavior in offspring, with a particular focus on sexual dimorphism. We analyzed 20 experimental studies involving rodent models (rats and mice) exposed to either lipopolysaccharide (LPS) or POLY I:C during gestation following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. Our findings reveal that most studies report a detrimental impact of MIA on the learning and memory performance of offspring, highlighting the significant role of prenatal environmental factors in neurodevelopment. Furthermore, this review underscores the complex effects of sex, with males often exhibiting more pronounced cognitive impairment compared to females. Notably, a small subset of studies report enhanced cognitive function following MIA, suggesting complex, context-dependent outcomes of prenatal immune challenges. This review also highlights sex differences caused by the effects of MIA in terms of cytokine responses, alterations in gene expression, and differences in microglial responses as factors that contribute to the cognitive outcomes observed


Maternal immune activation and congenital disorders in humans
Maternal immune activation during pregnancy is increasingly linked to long-term neurological and psychiatric disorders in offspring.In recent years, activation of the maternal immune system has been strongly associated with the development of specific cognitive and behavioral impairments in offspring (Ozaki et al., 2020;Vlasova et al., 2021).
The World Health Organization (WHO) emphasizes the severe impact of congenital disorders, which cause substantial neonatal and child mortality worldwide (World Health Organization, 2023).An estimated 240,000 newborns die annually within the first 28 days of life due to these disorders, with another 170,000 deaths occurring in children aged one month to five years.These conditions not only contribute to mortality but also to long-term disability, affecting individuals, families, and healthcare systems, particularly in low-income and middle-income countries (Ganguli and Chavali, 2021).
These data remain relatively limited even in less-developed countries, where the issue is more pronounced.This also explains why our knowledge of mental disorders and their relationship with maternal immune activation (MIA) is limited, although it does not excuse the lack of sufficient investigations addressing these issues.In addition, it is worth mentioning that in 2020 the COVID-19 pandemic had a devastating global impact, affecting millions of pregnant women.However, the full extent of its effects is still relatively unknown.The SARS-CoV-2 respiratory virus pandemic has provided a testing ground for some of the possible effects of MIA in humans.Research such as that of Zhu et al. (2020) has shown that neonates of infected mothers with a strong immune response to COVID-19 exhibited fever, abnormal liver function, and elevated heart rate, among other symptoms.Other studies revealed increased levels of IgM and IgG, as well as IL-6, in such patients (Dong et al., 2020;Riedel et al., 2022;Zeng et al., 2020).These are some immediate consequences, but their long-term consequences are currently unknown.However, more recent studies (Tetsuhara et al., 2022;Alves et al., 2022;Martin et al., 2021) have reported cases of SARS-CoV-2 infection in neonates, including seizures, abnormal brain magnetic resonance imaging (MRI), and other neurological signs (Brum and Vain, 2023).Therefore, examining this issue from this perspective is crucial.

Mechanisms of action of MIA on offspring
Studies on the behavioral and neurodevelopmental effects of MIA in rodent animal models date back to the beginning of the past century (Meyer and Feldon, 2009;Zuckerman et al., 2003).These researchers performed multiple studies that set the groundwork for research in this field.They delved deeper into some of the long-term neuropathological consequences of MIA, focusing on its impact on brain areas and neurotransmitter systems, as well as on behavioral outcomes, such as working memory or schizophrenia-related behaviors (Meyer and Feldon, 2009;Zuckerman et al., 2003;Meyer and Feldon, 2012).
There are several explanations regarding how this "indirect impact" of immunocompromised mothers affects their unborn offspring.In this regard, Albrecht et al. (2021) shed light on the crucial role of the placenta in the vertical transfer of maternal antibodies, specifically IgG, which are vital for neonatal immunity against infections in the early postnatal period.This transfer relies heavily on the health and function of the placenta, which influences the effectiveness of maternal immunization strategies aimed at enhancing neonatal immune defenses.Jennewein et al. (2017) explored the comprehensive programming of a newborn's immune system through the transfer of multiple immunological components via the placenta.This transfer includes not only antibodies but also inflammatory mediators and micronutrients, which collectively help shape the immune response of newborns.
Pregnancy induces significant physiological and immunological changes which are essential for maternal-fetal communication.However, these adaptations also increase susceptibility to infections.Vertically transmitted infections, such as those from the ToRCHS group (toxoplasmosis, rubella, cytomegalovirus, HIV, and syphilis), as well as herpes simplex virus, hepatitis viruses, parvovirus B19, and Zika virus, pose severe risks to both the mother and fetus.These infections can be transmitted transplacentally, during labor, or through breastfeeding, leading to adverse pregnancy outcomes such as low birth weight, stillbirth, and preterm labor (Megli and Coyne, 2022;Boushra and Farci, 2023;Kumar et al., 2022).Additionally, several studies have reported that immune activation due to these infections can cause alterations in learning and memory processes during adulthood.For instance, the Toxoplasma gondii infection has been linked to learning and memory impairments in rats (Daniels et al., 2015).
Vertical transmission mechanisms vary according to the pathogen.Toxoplasma gondii, rubella virus, and cytomegalovirus can be transmitted transplacentally, with risks increasing with gestational age or during birth/breastfeeding.The herpes simplex virus is mainly transmitted through active lesions during delivery.These infections can cross the placental barrier, leading to adverse outcomes, such as low birth weight and preterm labor (Boushra and Farci, 2023;Kumar et al., 2022).

Immunogenic agents for MIA in animal models
A wide array of diseases, ranging from Alzheimer's disease to multiple sclerosis, can potentially be attributed to disruptions in fetal brain development (Knuesel et al., 2014).However, due to their ability to mimic the behavioral and pathophysiological symptoms of conditions such as schizophrenia and autism, prenatal infections are commonly used as animal models for these disorders (Suvisaari et al., 1999;Brown and Derkits, 2010;Atladóttir et al., 2010;Goines et al., 2011).Some of the most commonly used models for mimicking the effects of MIA are based on the administration of lipopolysaccharide (LPS) and polyinosinic:polycytidylic acid (POLY I:C).
Lipopolysaccharides are components of the cell wall of gramnegative bacteria.This endotoxin has three regions: a backbone composed of highly acylated diglucosamine (lipid A), which is connected to a polysaccharide (antigen O) linked through a Kdo/heptose oligosaccharide core (Gorman and Golovanov, 2022).On the other hand, polyinosinic:polycytidylic acid is a synthetic ribonucleic acid polymer that mimics the structure of certain types of viral double-stranded RNA nucleic acids, alternating between inosines and cytidines.Although both activate the innate immune response, POLY I:C activates TLR3 receptors, and LPS activates TLR4 receptors.This involves some differences in terms of POLY I:C, leading to increased production of type I interferons (α and β), whereas LPS further increases the release of TNF-α, interleukin-1 (IL-1), and other inflammatory molecules (He et al., 2021).
Although both LPS and POLY I:C infections cause similar cellular and physiological reactions, such as fever response or cytokine induction, they may not have the same effects on CNS functions, such as memory or cognitive flexibility.Some long-term effects of MIA are dependent on the infection that causes prenatal inflammation.For example, exposure to POLY I:C has been linked to the development of neurodevelopmental issues (Bao et al., 2022).In contrast, MIA caused by LPS has been linked to renal (Vieira et al., 2018), intestinal (Elgin et al., 2019), and behavioral problems (Brown et al., 2017).

MIA and cognition
However, the mechanism underlying these effects remains unclear.Currently, more specific studies using animal models are ongoing to determine if the alterations are caused by the infection itself, maternal immune response, or maternal behavior during the infection.It is crucial to understand specific behavioral alterations to explain how prenatal infections might affect memory in different ways.Often used as models for disorders, such as schizophrenia and autism, the true impact on basic cognitive abilities remains unknown.Batinić et al. (2016) suggested that infection may induce mental disorders, such as schizophrenia, but not autism, which can only be differentiated by examining behavioral aspects independently.
Several of these studies indicate that the synaptic and intrinsic properties of neurons are impaired in MIA offspring.Alterations in synaptic plasticity, particularly long-term potentiation (LTP), which is essential for learning and memory, have been observed in the hippocampus (Ito et al., 2010;Oh-Nishi et al., 2010).The components of the neurotransmitter system are also affected in the MIA offspring.Additionally, impaired hippocampal neurogenesis and increased microglial density have been associated with abnormal learning and memory behaviors in prepubescent MIA offspring (Zhao et al., 2019;Couch et al., 2021).Cytokines, such as TNFα, IL-1β, and IL-6, play a crucial role in these processes, with IL-6 being particularly significant in mediating changes in neurogenesis (Smith et al., 2007).Gold et al. (2010) found that patients diagnosed with schizophrenia maintained their accuracy in a working memory test while experiencing a decrease in working memory capacity.Similarly, Barch and Smith (2008) and Barch et al. (2001) reported working memory impairments in schizophrenia patients, specifically in goal maintenance, interference control, and memory capacity.In contrast, autism-related memory issues have been linked to linguistic, social, and cognitive flexibility problems (Amodeo et al., 2019;Garcia-Valtanen et al., 2020;Haida et al., 2019).This suggests that prenatal infection models for autism and schizophrenia lack accuracy and ecological validity.
Several studies have used animal models to test the mechanisms that contribute to MIA-induced cognitive impairment.Woods et al. (2023) investigated the dysregulation of reelin signaling in a rat model of MIA and reported reduced reelin expression in the prenatal and adult prefrontal cortices.These findings are associated with memory impairment and altered NMDA function, resulting in inefficient long-term potentiation.Similarly, Sun et al. (2022) examined how abnormal proBDNF expression in POLY I:C-treated offspring in the hippocampal region impacts contextual memory, noting that the proBDNF/p75NTR pathway is a potential therapeutic target.In addition, Speers et al. (2021) explored hippocampal phase precession and theta sequences, which provide a mechanism for the sequential ordering of memory.They found that animals exposed to MIA exhibited disorganized ordering due to variability in the starting phase of precession.This suggests that multiple S. Sal-Sarria et al. simultaneous mechanisms could influence the possible cognitive impairment induced by MIA.

Sex differences in the effects of MIA
Significant variations in the impact of prenatal infections based on sex have been reported, although there is still considerable controversy regarding the underlying mechanisms and the sex that is predominantly influenced.Limited research on sexual differences in this context is growing, with recent studies increasingly incorporating both male and female animals to address this gap (Bhargava et al., 2021;Bölte et al., 2023;Hines, 2020).Despite these efforts, comprehensive data on the differential effects of MIA across sexes are lacking (Hall et al., 2023).
Most clinical studies have demonstrated sex-dependent differences in the prevalence and manifestation of mental disorders linked to MIA, such as schizophrenia.However, few studies have explored the underlying causes of these differences.The prevailing hypothesis attributes these sex-specific effects to hormonal differences, particularly the influence of sex hormones, such as estrogen and testosterone (Gogos et al., 2015).However, the precise underlying mechanisms remain unknown, and further research is needed to provide more specific and detailed explanations.
It is also crucial to acknowledge that sex differences in response to prenatal infections are evident at various biological and behavioral levels.For instance, males may exhibit greater vulnerability to certain cognitive deficits, whereas females may show relative resilience in certain areas but heightened susceptibility in others (Gogos et al., 2020).Recent studies have begun to explore these dimensions in more depth, but much remains to be investigated.
This systematic review aimed to assess how different variables related to MIA affect learning and memory in rodent offspring, with a special focus on sexual dimorphism.We hypothesized that the effect of MIA on the learning and memory abilities of rodent offspring is influenced by the interaction of multiple factors, including the sex of the offspring, type of immunogenic agent (LPS or POLY I:C) used to induce MIA, rodent species used in experimental models, and timing of MIA during gestation.Specifically, we expected male offspring to exhibit more pronounced cognitive impairments than female offspring, with variations in outcomes depending on the immunogenic agent, rodent species, and gestational timing of immune activation.

Review protocol
This study adhered rigorously to the guidelines set forth by the Preferred Reporting Items for Systematic Review and Meta-Analysis Protocol (PRISMA) 2020 (Page et al., 2021).The application of the PRISMA framework, a well-established and widely recognized protocol in the scientific community, ensured a systematic and transparent approach to the planning, execution, and reporting of the present systematic review.

Information sources
The information was collected through bibliographic searches conducted in the Scopus, PubMed, and PsycInfo databases, specifically including articles published in the last 10 years, from 2013 to 2023.The search included the following terms: ("poly(I:C)" OR lps OR "maternal immune activation" OR "maternal immune challenge") AND ("sex differences" OR "sexual dimorphism") AND behavior AND (rat OR mice).

Study selection and data collection process
First, all articles that were duplicated as a result of searching three different databases (Scopus, PubMed, and PsycInfo) were manually removed.Afterwards, a list of all eligible contributions was compiled.Finally, the three researchers collaborated to reach a consensus on selecting records for the systematic reviews, based on the requirements of the chosen topic.

Eligibility criteria
The strategy followed to conduct this systematic review involved selecting a specific sample, type of intervention, time of intervention, and specific behavioral tests.When possible, complementary biological measures of interest were also included.Therefore, only publications containing experimental research were considered.
Regarding animal models, only studies involving rodents (rats or mice) of different strains and including both sexes were considered to study potential sexual dimorphisms.Additionally, complementary studies that divided their sample and published two separate articles were also included.
The intervention consisted of one or several injections of LPS or POLY I:C during the prenatal period.Thus, injections were administered to pregnant females whose offspring were included in the study.It could be argued that the mothers were not truly part of the experiment given that in most cases, only data on their offspring were collected.However, in some cases, the behavior of the mothers was considered relevant for the research.
In this case, information was collected on the immunogenic agent (LPS or POLY I:C), the dose injected (mg/kg), the route of administration (subcutaneous, intraperitoneal, or tail vein), the number of injections, and the specific time of gestation at which they were administered.
Finally, only studies with behavioral memory or learning tests were included, although most also included other tests that have been considered.In addition, biological variables such as maternal serum cytokine levels, gene expression, microglial distribution, and neuroinflammation were also considered.
Therefore, the exclusion criteria were defined equally by all three researchers as those who did not meet the aforementioned requirements.To be precise, items were excluded if other species of animals were used, MIA occurred at other stages of development, both males and females were not included in the sample, or if no tests were made for memory or learning-related behavior tasks.

Risk of bias
The tool used to assess the risk of bias (RoB) of the studies included in this review was SYRCLE.This tool is an adaptation of the Cochrane RoB Tool for Animal Experimentation (Hooijmans et al., 2014).
The Bas SYRCLE tool was integrated with 10 entries divided into six fields: selection bias, performance bias, detection bias, attrition bias, reporting bias, and other biases.For this study, a scale of 0-1 was given to each of the entries, with a maximum of 10 points in total and a minimum of 0. Therefore, the resulting quality indices are 0-5 (low), 5-8 (medium), and 8-10 (high)."Low" (L) refers to a study with a significant risk of bias, and "High" (H) refers to a research experimental study with negligible or nonexistent RoB.
All three researchers who conducted this study independently assessed the risk of bias.The assessment was then shared to reach the final joint decision, as shown in Table 1.

Selection of studies
The search strategy is illustrated in the flow diagram found in Fig. 1.First, the respective searches yielded 301 articles, which were reduced to 169 after eliminating duplicates.Subsequently, articles that contained behavioral sections with results related to memory or learning tests were screened; 35 articles remained.After being read and analyzed, S. Sal-Sarria et al. 16 articles were identified from the original search, and 4 more were identified after reading the full articles and obtaining bibliographies from them.Twenty experimental studies met all the inclusion criteria for the present review.

Search results and risk of bias
Among the 20 studies included, 50 % (10/20) used mice, specifically the C57BL/6 strain, whereas the remaining 50 % used rats.The primary strains were Wistar, Sprague -Dawley, and Long Evans rats.
Regarding the infection methods, most studies, accounting for 75 % (15/20), used POLY I:C infections.Among these, 50 % (10/20) of the animals exposed to maternal immune activation (MIA) around gestational day 15 received the injection, while the others were exposed earlier, around gestational day 9 or GD12.Notably, Meehan et al. (2017) compared two groups with different durations of infection and tested whether the day of infection was significant.
With regard to the number of injections, 65 % (13/20) of the studies administered a single injection, whereas 30 % (6/20) involved two or three injections.Vojtechova et al. (2021) reported the administration of 6 injections of 1 mg/kg each.
The intraperitoneal route was the most commonly used method of administration in 55 % (11/20) of the studies, followed by the intravenous route via the tail vein in 45 % (9/20) of the studies.Only 5 % (1/ 20) of the studies used the subcutaneous route of administration (Vojtechova et al., 2021).
Regarding the risk of bias and methodological quality of the selected studies, 11 were classified as medium (M) (55 %), seven as high (H) (35 %), and the remaining two as low (L), with the lowest quality in the review (10 %).

Learning and memory
Most articles reported a negative impact of MIA on memory or learning, accounting for 70 % (14/20) of the included studies.In contrast, 10 % (2/20) of the articles reported improvements in learning (Nakamura et al., 2021;Zhao et al., 2021).However, 15 % (3/20) of the studies found no significant differences between the MIA groups and controls (Amodeo et al., 2019;Fernández De Cossío et al., 2021;Hui et al., 2018).Finally, in the article by Vojtechova et al., (2021), there were no relevant data because even the controls could not successfully complete the Carousel Maze test.
Research in which no impact of MIA was found used the T-Maze (Amodeo et al., 2019), Barnes Maze (Fernández de Cossío et al., 2021), and Novel and Spatial Recognition tasks (Hui et al., 2018).All three experiments were performed with C57BL/6 mice.Amodeo et al. (2019) and Hui et al. (2018)  Research showing that MIA has a positive effect on memory or learning (Nakamura et al., 2021;Zhao et al., 2021) also used C57BL/6 J mice and intraperitoneal POLY I:C administration.The dose used by Zhao et al. (2021) was greater than that used in most studies (20 mg/kg during GD 12).The dose used in Nakamura et al. (2021) was 5 mg/kg, although it was injected on 3 consecutive days, GD 13-14-15.
A total of 70 % of the studies reported negative effects of prenatal infection on memory and learning.Of these articles, the study by Batinić et al. (2016) used the lowest dosa, a single intraperitoneal injection of 0.1 mg/kg LPS on GDs 15 and 16 in Wistar rats.The dosage schedule in this study was very similar to that from Wischhof et al. (2015), with the only difference being the use of two daily doses instead of one.All other aspects remained unchanged.These two studies revealed deleterious effects on memory, assessed using the Morris water maze (Batinić et al., 2016) and the object recognition task (Wischhof et al., 2015).In addition, the doses used in the research of Schaafsma et al. (2017) did not exceed 0.25 mg/kg using LPS.C57BL/6 mice were intraperitoneally administered 0.25, 0.10, or 0.05 mg/kg GD15-16-17.Learning and memory were assessed using the T-maze.
The remaining studies that analyzed the impact of MIA on learning or memory used doses of approximately 4-5 mg/kg, and all used POLY I:C.There were small differences in all the other aspects, as shown in Table 1.
Regarding studies assessing memory, one behavioral variable appears in multiple studies: cognitive flexibility.Cognitive flexibility was included in 55 % (11/20) of the articles.In this regard, only one article reported memory improvement in groups in which mothers were administered POLY I:C during pregnancy (Zhao et al., 2021).However, this improvement was only reported in females.During the reversal-learning phase of the touchscreen test, females in the POLY I:C group required fewer trials and made fewer errors compared to their respective controls.(Table2).
Two studies found no differences between MIA and control groups (Gogos et al., 2020;Vorhees et al., 2015).Gogos et al. (2020) used Long Evans strain rats whose dams were injected via the tail with a 4 mg/kg dose of POLY I:C on GD day 15.The task they performed to assess cognitive flexibility was the pairwise discrimination and reversal learning task, and they found no significant effect of MIA on either sex.Vorhees et al. (2015) used Sprague -Dawley rats whose dams were treated with a relatively high intraperitoneal dose of 8 mg/kg POLY I:C from GD14 to GD18.Behavioral assessments were conducted using the Morris water maze, and no differences were found between the MIA groups and controls in the reversal or shift probe trials.

Immune activation, gene expression and additional biological variables
Most of the selected studies measured immune activation.There was a consistent finding of an increased cytokine response, with 45 % (9/20) of the studies reporting increased cytokine activity.The remaining articles simply evaluated immune activation and assessed other related aspects, such as microglial response, expression of specific genes, or parvalbumin levels in the brain regions of interest.None of the reviewed articles reported the absence of differences in cytokine responses between the MIA and control groups.
However, regarding more basic variables, such as body weight, rectal temperature, and gestational length, not all studies reached consistent conclusions.While most studies did not indicate any significant changes in offspring body weight, temperature, or gestational length, Gogos et al. (2020) found that dams treated with POLY I:C exhibited decreased body temperature.Lins et al. (2018), ( 2019) reported a decrease in body weight and an increase in body temperature (Murray et al., 2017).O' Leary et al. (2014) reported shorter gestational lengths, and Mueller et al. (2018) also reported spontaneous abortions in the MIA group.However, Murray et al. (2017) reported an increase in gestational length.It is worth noting that not all the studies included these measures.
Furthermore, Vojtechova et al. (2021) used more specific anatomical measurements, particularly brain size.Both males and females in the MIA group exhibited enlarged brains, with the effect being more pronounced in males.This study also revealed enlarged brains and a reduction in parvalbumin-positive interneurons in the frontal cortex, with juvenile males displaying larger brains than females.Amodeo et al. (2019) reported the long-term effects of MIA on the expression of 24 genes, some of which were related to glutamatergic neurotransmission, potassium ion channel activity, and mTOR signaling.Hui et al. (2018) reported an increased expression of inflammation-related genes in the cerebral cortex and hippocampus of males.Nakamura et al. (2021) reported changes in Gad1 and long-term changes in hippocampal PV mRNA expression in adults.
In a study by Richetto et al. (2013), there was a significant reduction in the gene expression of GAD65 and GAD67 in the adult medial prefrontal cortex and dorsal hippocampus.Subsequently, Richetto et al. (2014) revealed altered mRNA expression patterns, consistent with deficits in prefrontal GAD65/GAD67 and VGAT expression at the protein level.The prefrontal GABAergic transcriptome, particularly in adults, exhibited a more significant impact, emphasizing the potential enduring consequences of prenatal immune activation on GABA-related mechanisms in the brain.
In another study by Zhao et al. (2021), sex-specific patterns related to the upregulation of genes essential for cognitive functioning were observed.Notably, significant reductions in CamK2a mRNA levels, specifically in the prefrontal cortex, were detected.
The  (continued on next page) (2021).A reduction in the microglial response during adulthood was found in a study by Schaafsma et al. (2017).In contrast, Hui et al. (2018) reported a significant increase in the density of dark microglia in the dentate gyrus, particularly in males but also in females.Finally, Vojtechova et al. ( 2021) reported no differences in the number of microglia.Fernández de Cossío et al. ( 2021) also failed to find differences in the number of microglia in the dorsal hippocampus measured by cytometry at postnatal day 15.They also reported a greater number of microglia in the dentate gyrus in the male group compared to the female group.Finally, 20 % (4/20) of the investigations specifically explored parvalbumins in depth: Vojtechova et al. (2021), Nakamura et al. (2021), Wischhof et al. (2015) and Zhao et al. (2021).Zhao et al. (2021) and Nakamura et al. (2021) reported increased numbers of PV cells in the prefrontal cortex and hippocampus, respectively.

Sex differences
Although all the studies mentioned including both males and females in the experiment and even discussed possible sexual dimorphisms, not all of them showed results in this regard.In 20 % (5/20) of the studies, females did not undergo the same evaluation as males did (Amodeo et al., 2019;Murray et al., 2017;Richetto et al., 2014;Richetto et al., 2013;Schaafsma et al., 2017).
Amodeo et al. ( 2019) focused exclusively on males, omitting any mention of females.Murray et al. (2017) initially included both males and females but later narrowed their focus only to male subjects.Richetto et al. ( 2013) conducted behavioral tests with mixed-sex groups due to the absence of significant sex differences, but molecular tests exclusively considered males.In the study by Richetto et al. (2014), only males were included in both cognitive and behavioral assessments 'to circumvent bias arising from sexual dimorphism'.Finally, Schaafsma et al. (2017) included females in some basic measures, such as body weight, but did not include them in other tasks or measurements.
In almost all cases, males showed deficits, or in cases where both sexes showed deficits, males were more affected.This was also observed in a study conducted by Batinić et al. (2016), who reported that while males showed impaired memory, females did not.Additionally, amphetamine-induced locomotion was significantly lower in the female offspring of the LPS-treated dams than in the male offspring.Meehan et al. (2017) reported reduced prepulse inhibition (PPI) in MIA males, along with increased dopamine type 1 receptor mRNA levels in the nucleus accumbens, but not in females.Lins et al. (2018) found increases in locomotion and reduced sociability only in males.Vorhees et al. (2015) noted that males exhibited more freezing behavior than females under conditions of fear/latent inhibition.Similarly, Wischhof et al. (2015) reported more severe consequences in various aspects in females, although they were also present in males to a lesser extent.Hui et al. (2018) revealed a more pronounced impact of POLY I:C on behavioral aspects related to anxiety and sensorimotor synchronization in males compared to females.In a study of microglia, alterations in the number of contacts with myelinated axons were also found in females, whereas in males, there was an increase in the area and perimeter of the microglial process.
In contrast, O'Leary et al. ( 2014) reported a decline in spatial working memory in female offspring but not in male offspring.They also noted that PPI was more strongly affected in females than in males, especially in POLY I:C-treated dams.Zhao et al. (2021) also observed sexual dimorphism.However, in this case, it was related to the benefits obtained from MIA.The female offspring of POLY I:C-treated dams exhibited better learning abilities compared to the male offspring under the same conditions.Similarly, Nakamura et al. (2021) reported sex differences in TUNL performance, indicating better working memory in males in this task.
Finally, Vojtechova et al. (2021) mentioned several sex differences, including a larger brain size in males compared to females, as well as a greater number of parvalbumin-positive (PV+) interneurons in the upper part of the frontal cortex but a lower number of PV+ interneurons in the dorsal hippocampus.Regarding behavioral evaluation, males showed deficits in communication and social behavior earlier than females did, whereas females exhibited more anxious behaviors.However, females showed PPI deficits, whereas males did not.

Discussion
The systematic review of studies examining the impact of maternal immune activation on learning and memory in rodent models revealed several critical findings.The overarching narrative emphasizes how prenatal infections or immune challenges can impact neurodevelopmental outcomes, specifically behavioral outcomes such as learning and memory.Furthermore, this review highlights the relevance of sexual dimorphism to these effects, offering a more comprehensive understanding of how prenatal environmental factors shape neurobehavioral trajectories in a sex-dependent manner.

Behavioral and immune alterations
The collective body of research reviewed reveals a predominant trend in which MIA is associated with a detrimental effect on the learning and memory ability of rodent offspring.Specifically, 70 % (14 out of 20) of the studies reported negative outcomes in memory or learning following prenatal exposure to immune challenges.These findings are supported by the studies from Batinić et al. (2016) and Wischhof et al. (2015), who illustrated how exposure to LPS or POLY I:C during gestation can impair cognitive function in offspring, which was assessed through behavioral tasks such as the Morris water maze and object recognition tasks during adulthood.
The assertion that MIA can impair cognitive function in rodent offspring is robustly supported by a broader spectrum of studies than those initially cited.Kirsten et al. (2013), Yin et al. (2013), and Lanté et al. (2007), (2008) found similar spatial learning and memory deficits in male rodents exposed to several LPS doses during late gestation.Synaptic alterations in the hippocampus due to prenatal LPS exposure, such as decreased presynaptic input to the CA1 layer and compensatory enhancement of pyramidal cell excitability (Lowe et al., 2008), may explain these findings.These alterations can disrupt the development and function of the hippocampal circuitry, which is crucial for spatial learning and memory, as evidenced by the changes in neuronal composition and function reported by Baharnoori et al. (2009) and Cui et al. (2009).
Remarkably, sex differences were found in the cognitive function of the offspring after MIA.A recurrent finding in numerous studies is the differential effect of prenatal immune challenges on the neurodevelopment of male and female offspring.However, adversity is often more pronounced in males.This sex-specific response to MIA is evidenced by the more pronounced cognitive impairment, particularly in learning and memory, observed in male rodents compared to their female counterparts.For instance, Batinić et al., 2016 andMeehan et al., 2017 highlighted that male rodents typically display greater cognitive deficits after MIA.This observation is further supported by Wischhof et al. (2015), who found that male rats subjected to prenatal LPS challenge showed significant memory impairment in the Morris water maze and object recognition tasks.Similarly, Schaafsma et al. (2017) observed that male mice exposed to LPS demonstrated reduced performance in the T-maze, which is indicative of learning and memory challenges.These studies collectively highlight the variable impact of prenatal immune challenges based on the sex of the offspring, suggesting a critical need for further research on the mechanisms driving these sex-specific neurodevelopmental outcomes.However, it is important to acknowledge that the picture of sexual dimorphism is complex and not uniformly skewed toward male vulnerability.Some studies, such as those by O' Leary et al. (2014) and Zhao et al. (2021), have shown that female rodents experienced cognitive benefits from MIA or were more adversely affected in certain cognitive domains.
Many studies have highlighted the role of inflammation-related factors, particularly maternal interleukin-6 (IL-6) elevation, which can lead to lasting neurobiological changes in the fetus (Boksa, 2010;Boulanger et al., 2018;Canetta et al., 2014;Coulthard et al., 2018).Although other inflammation-related factors, such as IL-1α, TNF-α, and IFNγ, are inducers of IL-6 in vivo, they do not appear to be decisive in this regard (Gadient and Otten, 1997;Smith et al., 2007).This association has been tested in human research, such as in the study by Rudolph et al. (2018), in which aspects such as emotionality and working memory were measured, and no differences were found.The measurement of working memory involved a task in which children had to recall the locations of stickers hidden in six of eight pots, termed the spin-the-pot task.An association was observed between elevated IL-6 levels and reduced performance scores in this task.
Building on the role of IL-6, recent findings, such as those reported by Posillico et al. (2021), emphasize the existence of sex-specific neuroimmune responses to the central administration of POLY I:C.This study sheds light on foundational neuroimmune differences and similarities between males and females, which could underlie the unique compensatory mechanisms observed in females following MIA.They found that both male and female mice exhibited weight loss, fever, and elevated levels of cytokines and chemokines in the hippocampus following the intracerebroventricular administration of POLY I:C.However, the response was generally stronger in males than in females, with males showing increased levels of IFNα and IFNγ, which are markers that were not elevated in females.In contrast, both sexes exhibited elevated IFNβ levels, indicating a sex-specific difference in the antiviral response within the hippocampus.These findings suggest that type I interferons could be a crucial factor mediating sex-specific cytokine responses, affecting the neuroimmune effects on cognition.Carlezon et al. (2019) reported that prenatal exposure to POLY I:C or LPS resulted in more pronounced behavioral impairments that align closely with the core features of autism spectrum disorders, such as altered social behavior and increased anxiety-like behaviors, which were not as evident or were entirely absent in females.Sex differences extend to molecular responses, with significant differences observed in the expression of proinflammatory and neuroinflammatory markers, where changes were generally more substantial in males than in females.Remarkably, this review identified contrasting patterns of changes in anti-inflammatory markers between sexes following perinatal immune activation.Specifically, while both IL-10 and TGF-β1 were downregulated in males, indicating a proinflammatory state, these markers were significantly upregulated in females, suggesting a potential anti-inflammatory or protective response that could mitigate the severity of behavioral phenotypes in female mice.
Research performed in rodents has shown that IL-6 from the mother can pass through the placenta and reach the fetus in the middle stages of gestation but not in the later stages, according to Dahlgren et al. (2006).As pregnancy progresses, the placenta becomes crucial for mediating the effects of maternal immune activation, leading to inflammation in both the placenta and the fetal brain.Wu et al. (2017) reported that removing IL-6 receptors from placental trophoblasts in mice can prevent adverse outcomes triggered by prenatal infection.In this regard, it seems that males are more susceptible to the effects of MIA, probably because of the protective effect of endogenous estrogens in females.In addition, the sexual differentiation of specific brain areas and behaviors involves components of signal transduction pathways common to those involved in inflammatory processes (Taylor et al., 2012).Thus, immune effects may lead to sex-related alterations, such as those found in this review.However, it is still unclear how the rise in interleukin or alterations in sexual development due to MIA result in changes to the typical neurodevelopment of the offspring.
In contrast, a smaller fraction of the studies, notably 10 % (2 of 20), documented an improvement in learning or memory following MIA.Nakamura et al. (2021) and Zhao et al. (2021) reported that certain doses of POLY I:C could enhance memory performance in C57BL/6 J mice.However, these outcomes appear to be exceptions rather than rules, highlighting the need for further exploration of the conditions under which MIA might lead to cognitive benefits.
The potential cognitive benefits of MIA, although less common, deserve consideration due to their significance in understanding the intricate nature of prenatal environmental effects.Smith et al. (2007) explored the hypothesis that low-level immune challenges during pregnancy could, in certain contexts, are not enough to impact long-term behavior.The immune challenge must be significant and prolonged over time, as in influenza infection (Yang and Evans, 1961;Shi et al., 2003), leading to a considerable increase in maternal IL-6 and, subsequently, to changes in the behavior of the offspring.In this regard, Smith et al. (2007) showed that when IL-6 is removed from the maternal immune response, either through genetic modification or through the use of blocking antibodies, adult offspring do not exhibit behavioral deficits typically associated with MIA.These results are similar to those recently reported by Wu et al. (2017).Additionally, Makinson et al. S. Sal-Sarria et al. (2019) reported findings that align with the observations of Nakamura et al. (2021) and Zhao et al. (2021), showing that exposure to immunogenic agents during gestation can result in enhanced cognitive abilities.In a study by Makinson et al. (2019), MIA improved cognitive performance in the 5-choice serial reaction time task.Zhao et al. (2021) referred to the proposal of "hidden talent" by Ellis et al. (2022), whereby subjects exposed to harsh and unpredictable environments during childhood may develop compensatory mechanisms and present some cognitive advantages later in life.However, it is important to emphasize that the mechanisms by which MIA can lead to benefits are unclear.The scientific literature documenting the benefits of MIA is almost non-existent, being Zhao et al. (2021) or Nakamura et al. (2021) among those reporting beneficial effects.
Nevertheless, it is crucial to highlight that only a limited number of studies have demonstrated the beneficial effects of MIA on cognitive function.Moreover, a consistent observation is that the impact of MIA is less severe in females than in males.This does not imply that females benefit from it, but rather that they are more resilient to its effects.
The impact of MIA on cognitive function extends to cognitive flexibility, a critical component of executive functioning that involves the ability to switch between different concepts or adapt to changing environments.Cognitive flexibility is fundamental for learning from past experiences and applying knowledge in novel contexts, making it crucial to understand the full spectrum of neurodevelopmental outcomes following prenatal immune challenges.
A significant number of the studies in this review reflect general research trends and indicate that prenatal exposure to immune challenges, such as LPS or POLY I:C, can result in cognitive flexibility impairments.This finding aligns with observations in the general literature, indicating that MIA is associated with disruptions in brain regions crucial for executive function, particularly in the prefrontal cortex.Studies such as those by Batinić et al. (2016) and Meehan et al. (2016) provide evidence of such impairments, with male rodents often showing more significant deficits post-MIA than females, suggesting sex-dependent vulnerability.
The study by Arsenault et al. (2014) provides a comparative analysis of the behavioral and developmental effects of LPS and POLY I:C exposure during pregnancy.Their findings revealed distinct patterns of anxiety-like behaviors in dams and developmental delays in offspring, further supporting the association between prenatal immune challenges and subsequent impairment in cognitive flexibility.Cognitive flexibility, a complex cognitive ability, can be altered directly or indirectly by other more fundamental factors, such as the modulating role of anxiety in attention.Arsenault et al. (2014) also showed that POLY I:C triggered a delay in growth and sensorimotor development, which is also related to cognitive flexibility, as this has much to do with the prefrontal cortex, a brain region that requires a significant amount of time to fully develop.
A study conducted by Haddad et al. (2020) explored the mechanisms by which MIA affects the prefrontal cortex and revealed that inflammatory cytokines, particularly IL-6, play a pivotal role in these effects.Spann et al. (2018) added to these findings by demonstrating that MIA during the third trimester correlates with altered functional connectivity in the salience network of newborns, which includes regions such as the medial prefrontal cortex.This study highlights that elevated levels of inflammatory cytokines such as IL-6 and C-reactive protein during pregnancy are linked to altered functional connectivity in the medial prefrontal cortex, temporoparietal junction, and basal ganglia of neonates.Furthermore, this dysregulation of neural networks can result in lasting cognitive and behavioral issues, reinforcing the link between prenatal immune challenges and cognitive impairments identified in previous studies included in this review.
Several studies have reported sex-based differences in the effects of MIA on cognitive flexibility (Gogos et al., 2020).Once again, males were the most affected.The effects of estrogen on spine density in the female hippocampus (MacLusky et al., 2005) may explain the differences in hippocampus-dependent tasks related to cognitive flexibility, such as a delayed-to-location task using a touchscreen (McAllister et al., 2013).
Despite extensive efforts to explain the mechanisms involved in sex differences in learning and memory following MIA, including studies with methodological errors, we still lack scientific certainty to explain the biological mechanisms underlying the modulator role of sex in how MIA affects learning and memory.

Timing effects of MIA and rodent species/strains used
Our review also showed that the gestational day of exposure to MIA plays a role in the memory and learning impairment observed in offspring.Notably, exposure during the mid-to late-gestation period (GD 15-17) was more consistently associated with significant memory deficits than early exposure (GD 9-12).For instance, studies such as those by Batinić et al. (2016) and Gogos et al. (2020) reported impaired memory after MIA on GD 15, whereas early exposures such as in the study by O'Leary et al. ( 2014) (GD 9) also showed impaired memory effects, although this finding was not consistent.Therefore, our results support the findings of previous studies suggesting that exposure to MIA during mid-late gestation (after GD 14-15) in rodent models of schizophrenia predominantly results in memory impairments and social withdrawal compared to earlier MIA exposure (GD 9-10), which is associated with deficits in prepulse inhibition and amphetamine sensitivity (Gray et al., 2019).These differences in the timing of MIA exposure could be related to critical periods of brain development, such as the proliferation and migration of limbic cortical neurons, particularly synaptogenesis and peak neurogenesis in the hippocampus between GD 14 and 17 (Semple et al., 2013;Zuckerman et al., 2003;Rice and Barone, 2000).
The frequency of immunogenic agent administration also significantly influences patient outcomes.Exposures extending beyond a single day, such as those in Batinić et al. (2016) (2 days) and Vorhees et al. (2015) (5 days), tend to show more pronounced and consistent effects on decreased memory and cognitive flexibility.This suggests that both the timing and duration of exposure are critical in determining the degree of cognitive impairment in offspring.
The type of animal species and strain used in these studies also contributed to the observed variation.Different animal strains may have varying susceptibilities to MIA, which could explain the diverse results.For example, some strains might be more resistant or susceptible to the inflammatory effects of immunogenic agents, such as LPS or POLY I:C, subsequently affecting memory and learning outcomes.In this regard, it is important to mention that only two studies that reported the beneficial effects of MIA (induced by POLY I:C) were performed with C57BL/ 6 J mice (Nakamura et al., 2021;Zhao et al., 2021).
In summary, the synthesis of findings from this systematic review sheds light on the intricate ways in which prenatal environmental factors, such as maternal immune activation, can influence neurodevelopmental trajectories in offspring.The overwhelming evidence reviewed here suggests a negative effect of MIA on learning and memory in rodent models, with an interesting layer of complexity added by the observed sexual dimorphism.These insights not only advance our understanding of the biological underpinnings of neurodevelopmental disorders but also underscore the need for further research that considers sex as a critical variable in the study of prenatal environmental exposures.

Limitations
This review may be biased by primarily including published studies with significant results and potentially overlooking unpublished or negative data.This could overestimate the effects of maternal immune activation on cognitive outcomes.The variability in study design, methods, and assessments introduces heterogeneity, complicating comparisons and limiting generalizability.Additionally, concerns about sample sizes for detecting sex-dependent effects, variations in rodent strain susceptibility, and the focus on short-term studies may limit the broader applicability of these findings.

Conclusions
Despite these limitations, the review consistently shows that MIA detrimentally impacts cognitive functions in rodent offspring, as evidenced by impaired learning and memory.These effects are mediated by cytokine pathways, with maternal IL-6 elevation playing a pivotal role.This provides insights into the neurobiological alterations induced by MIA that mirror the symptoms of human neurodevelopmental disorders.Interestingly, a subset of studies suggests that mild prenatal immune challenges might present "hidden talents", enhancing cognitive performance under certain conditions and hinting at the complex, contextdependent impact of MIA.
Moreover, this review highlights cognitive flexibility impairments due to MIA and emphasizes sexual dimorphism in neurodevelopmental outcomes, with males being more affected and female outcomes changing depending on specific cognitive domains.This highlights the importance of considering sex in neurodevelopmental research, focusing on biological differences and prenatal insult specifics.
To advance our understanding, future research should investigate the mechanisms underlying sex-specific effects of MIA.This can be achieved through RNA sequencing and proteomic analyses to compare gene expression and protein profiles between males and females in MIA models.In addition, it is necessary to continue examining the long-term impact of MIA on cognitive and behavioral functions.The role of sex differences in microglial density, morphology, and activation in relation to synaptic pruning and maturation during brain development needs to be addressed (Hall et al., 2023).Furthermore, exploring the interactions between prenatal and postnatal factors, such as the social environment and stress, can provide insights into how these factors collectively influence neurobehavioral development.Experimental designs should include controlled variations in the postnatal environments of animals exposed to MIA.Arguably, the advantages of MIA in some of the papers studied here are related more to the postnatal environment than to the infection itself.In addition, different time windows of exposure to MIA or rodent species should be explored given that very early or very late prenatal stages have not been analyzed in detail.This may lead to interventions mitigating the adverse effects of MIA and exploring conditions where prenatal immune challenges could benefit cognitive development (if possible).These insights are crucial for unraveling the biological foundation of neurodevelopmental disorders and emphasizing the significant impact of prenatal environmental factors on shaping neurobehavioral outcomes.

Financial support
This work was supported by the Spanish Ministry of Science and Innovation grants PID2022-140980NB-I00 to HGP and NMC and the "Severo Ochoa" Program of Predoctoral Grants of the Regional Council of Culture and Sports of the Principality of Asturias (PA-22-BP21-020) to SSS.

Declaration of Competing Interest
None.
used POLY I:C as an immunogenic agent, while Fernández De Cossío et al. (2021) used LPS.The intraperitoneal route of administration was the same as that used in the previous three studies.There were no matches between the studies regarding the time of administration and dose used.Amodeo et al. (2019) used 20 mg/kg on GD 12.5, Fernández De Cossío et al. (2021) administered 0.1 mg/kg on GD 15, and Hui et al. (2018) administered 5 mg/kg on GD 9.5.

Fig. 1 .
Fig. 1.PRISMA flow diagram of the reviewing processsystematic selection for inclusion or exclusion.

Table 1
Description of research included.

Table 2
Behavioral and other results.