Global research priorities for infections that affect the nervous system

Infections that cause significant nervous system morbidity globally include viral (for example, HIV, rabies, Japanese encephalitis virus, herpes simplex virus, varicella zoster virus, cytomegalovirus, dengue virus and chikungunya virus), bacterial (for example, tuberculosis, syphilis, bacterial meningitis and sepsis), fungal (for example, cryptococcal meningitis) and parasitic (for example, malaria, neurocysticercosis, neuroschistosomiasis and soil-transmitted helminths) infections. The neurological, cognitive, behavioural or mental health problems caused by the infections probably affect millions of children and adults in low- and middle-income countries. However, precise estimates of morbidity are lacking for most infections, and there is limited information on the pathogenesis of nervous system injury in these infections. Key research priorities for infection-related nervous system morbidity include accurate estimates of disease burden; point-of-care assays for infection diagnosis; improved tools for the assessment of neurological, cognitive and mental health impairment; vaccines and other interventions for preventing infections; improved understanding of the pathogenesis of nervous system disease in these infections; more effective methods to treat and prevent nervous system sequelae; operations research to implement known effective interventions; and improved methods of rehabilitation. Research in these areas, accompanied by efforts to implement promising technologies and therapies, could substantially decrease the morbidity and mortality of infections affecting the nervous system in low- and middle-income countries.

Infections that cause significant nervous system morbidity globally include viral (for example, HIV, rabies, Japanese encephalitis virus, herpes simplex virus, varicella zoster virus, cytomegalovirus, dengue virus and chikungunya virus), bacterial (for example, tuberculosis, syphilis, bacterial meningitis and sepsis), fungal (for example, cryptococcal meningitis) and parasitic (for example, malaria, neurocysticercosis, neuroschistosomiasis and soil-transmitted helminths) infections. The neurological, cognitive, behavioural or mental health problems caused by the infections probably affect millions of children and adults in low-and middle-income countries. However, precise estimates of morbidity are lacking for most infections, and there is limited information on the pathogenesis of nervous system injury in these infections. Key research priorities for infection-related nervous system morbidity include accurate estimates of disease burden; point-of-care assays for infection diagnosis; improved tools for the assessment of neurological, cognitive and mental health impairment; vaccines and other interventions for preventing infections; improved understanding of the pathogenesis of nervous system disease in these infections; more effective methods to treat and prevent nervous system sequelae; operations research to implement known effective interventions; and improved methods of rehabilitation. Research in these areas, accompanied by efforts to implement promising technologies and therapies, could substantially decrease the morbidity and mortality of infections affecting the nervous system in low-and middle-income countries.
Nature 527, S178-S186 (19 November 2015), DOI: 10.1038/nature16033 This article has not been written or reviewed by Nature editors. Nature accepts no responsibility for the accuracy of the information provided.
R ecent improvements in the detection of infectious organisms that can affect the nervous system has led to the realization that a substantial proportion of chronic neurological, cognitive and behavioural disease may actually have an acute and preventable origin. Infectious organisms may infect the nervous system directly, as in rabies and bacterial meningitis, or may cause neurocognitive disorders in the absence of direct infection of the nervous system, as in malaria or hookworm infection. This Review identifies global research priorities for infections that affect the nervous system, with the ultimate goal of stimulating research in these priority areas to substantially reduce morbidity associated with nervous system infections worldwide.

METHODS
For this Review, we chose illustrative infections that cause considerable nervous system morbidity in children and adults in low-and middle-income countries (LMICs). These infections are examples and are not meant to be exhaustive. Estimates of infection global frequency and types of nervous system involvement were obtained through Pu-bMed searches using the infection name and were accompanied by any of the following terms: neurologic, nervous system, cognition, cognitive, development, neurodevelopment, impairment, deficit, sequelae, brain injury, brain damage, mental health, behavioral or neuropathy. If available, World Health Organization (WHO) documents were also reviewed for each disease. The authors came to a consensus on the key research priority areas, on the basis of a literature review and research experience. S179 present in as many as 70% of those with HSV encephalitis 4 and 30-50% of those with JEV encephalitis 5 . Most cases of JEV infection (as opposed to encephalitis) are asymptomatic or mildly symptomatic and require no treatment 6 , but the long-term neurocognitive consequences of asymptomatic or mildly symptomatic JEV infections are unknown.
Varicella zoster virus (VZV), the cause of chickenpox, can, like other herpesviruses, establish a latent infection. Reactivation is usually owing to suppression of cell-mediated immunity, most commonly age-related immunosenescence. Central nervous system (CNS) reactivation is relatively uncommon, but reactivation in a dorsal root ganglion can lead to herpes zoster, which is associated with debilitating chronic pain 7 . Herpes zoster is the single most common infection of the nervous system, with an estimated one million new cases each year in the United States alone 8 . There is little data on nervous system VZV infection in LMICs.
Congenital cytomegalovirus (CMV) infection is the most common acquired cause of hearing loss in children in the United States 9 . The percentage of the population testing positive for CMV is higher in LMICs than in high-income countries, but the incidence of congenital CMV infection and of symptomatic CNS disease and hearing loss in most LMICs remains almost unknown 10 .
Dengue, and to a lesser extent chikungunya, viruses will probably become leading global causes of arboviral encephalitis in the next decade 2,11,12 . Although encephalitis is present in only a fraction of those infected with dengue virus, the large number of dengue infections worldwide could lead to hundreds of thousands of encephalitis cases (Table 1).

HIV and opportunistic infections
In 2012, an estimated 2.3 million people were infected with HIV, and 1.6 million died of AIDS-related illnesses worldwide 13 . HIV-associated neurological syndromes are classified as primary HIV infection, secondary or opportunistic infection, and treatment-related neurological disease. Most often, primary HIV infection causes acute aseptic (viral) meningitis or meningoencephalitis (MEC). HIV-associated neurocognitive disorder (HAND) is a neurodegenerative condition characterized by cognitive, motor and behavioural abnormalities that is becoming more common with an increase of HIV in people older than 50 years 14 .
In LMICs, where only one third of patients requiring highly active antiretroviral therapy (HAART) receive it, the opportunistic infections cryptococcal meningitis, tuberculous meningitis, cerebral toxoplasmosis, progressive multifocal leukoencephalopathy and CNS cytomegalovirus infection remain common 15 .
The widespread implementation of combination antiretroviral therapy (cART) has changed the presentation, manifestation and epidemiology of many conditions and opportunistic infections, owing to immune reconstitution syndrome and increased prevalence of cognitive impairment and neuropathy with additional morbidities in patients who are now living longer 16 . The epidemiology and neurological outcomes of HIV infection are also affected by underlying malnutrition and variations in endemic pathogens.
Cryptococcal meningitis is a leading cause of mortality in LMICs where access to antiretroviral therapy is limited 17 . Most cases occur in sub-Saharan Africa, followed by South and Southeast Asia 18 . Seroprevalence for Toxoplasma gondii in people with HIV ranges from 10% to 80% with the highest proportions in African countries 19 , and cerebral toxoplasmosis is the most common cerebral mass lesion in patients with AIDS.

Bacterial infections
The most common bacterial infections affecting the nervous system are sepsis and meningitis in neonates; bacterial meningitis due to Streptococcus pneumoniae, Haemophilus influenzae type b and Neisseria meningitidis in children and adults; and tuberculous meningitis in children and adults.
Neonatal meningitis and neonatal sepsis are associated with longterm neurological and cognitive impairment 20 ; primarily impairment of hearing, vision or motor function; cerebral palsy; and epilepsy. In LMICs, it is estimated that 23% of neonates who have survived meningitis sustain moderate to severe neurodevelopmental impairment 21 . Staphylococcus aureus, Gram-negative infections, including Escherichia coli, Klebsiella pneumoniae, Acinetobacter, non-typhoidal Salmonella and group B Streptococcus, are the leading causes of neonatal sepsis and meningitis in most LMICs [22][23][24] . Recent reports from Africa and India suggest an alarming increase in drug resistance among Gram-negative organisms infecting neonates 24,25 .
In children and adults in high-income countries, bacterial meningitis due to S. pneumoniae, H. influenzae type b and N. meningitidis has decreased dramatically following immunization with conjugate vaccines 26 . However, availability of these vaccines in LMICs is variable, and bacterial meningitis still affects 1.2 million individuals annually 26 , causing neurocognitive sequelae in 23% of affected children. Steroids as an adjunctive therapy have reduced neurological sequelae in high-income countries, particularly in adults, but have shown no benefit in LMICs 27 . Factors such as organism strain, co-infections, adjunctive and supportive therapies, and underlying conditions such as poor nutrition that can affect immune response often differ between high-income countries and LMICs, and may play a part in the different results seen in some clinical trials.
Tuberculous meningitis (TBM) occurs in around 1% of all cases of tuberculosis, but is the most severe form of extrapulmonary tuberculosis, resulting in death or severe disability in about 50% of those with the disease 28 . Bacterial meningitis, and particularly TBM, can result in hydrocephalus, which is often difficult to treat in LMICs because neurosurgeons are typically not available and the supplies for ventriculoperitoneal shunt placement are often not present or very limited.
The WHO estimated that approximately 10.6 million new cases of syphilis occurred in 2008 (ref. 29), but precise estimates of the incidence of neurosyphilis are not available.

Parasitic infections
Malaria in humans is caused by one of five Plasmodium species, but neurological disabilities are most frequently associated with Plasmodium falciparum. Although P. falciparum does not directly infect brain tissue, severe infection can lead to coma. One in four children with cerebral malaria develops long-term cognitive impairment 30 , and recent studies suggest that children with severe malarial anaemia also have long-term cognitive impairment 31 . Behavioural problems and epilepsy are other long-term consequences of cerebral malaria 32 . Children with repeated episodes of uncomplicated malaria have motor and cognitive problems 33 . The mechanisms by which malaria leads to neurocognitive problems are not fully defined, and the neurocognitive burden of malaria owing to other Plasmodium species has not been characterized.
Neurocysticercosis, which is endemic in areas with poor pig management practices and sanitation, occurs when the larval stages of Taenia solium infect the brain. In LMICs in which neurocysticercosis is endemic, it is the leading identified cause of seizures. The proportion of larval infections migrating to the brain is unknown, but some individuals with neurocysticercosis never show neurological symptoms 34 . Neurocysticercosis can be fatal, most often following complications of surgery to treat the hydrocephalus associated with intraventricular or subarachnoid neurocysticercosis, but overall mortality estimates have been difficult to obtain 35 .
Neuroschistosomiasis results from the migration of schistosome eggs or worms to the brain or spinal cord, and may occur following infection with Schistosoma japonicum, Schistosoma mansoni or Schistosoma haematobium. Brain involvement may occur in the acute phase (acute schistosomal encephalopathy) or in the chronic phase (cerebral schistosomiasis or pseudotumoral encephalic schistosomiasis). The spinal cord may also be involved, often leading to hemiparesis 36 . CNS symptoms and epilepsy are reported to occur in 2.6% and 2.1% of S. japonicum infections, respectively 37 . Neuroschistosomiasis can be fatal, especially in its tumour-like form when it affects the cerebellum, but accurate mortality rates are unavailable 38 .  Soil-transmitted helminths (STH) affect millions, most frequently children. Determining the part played by STH in cognitive impairment of children is complicated because STH infection is associated with many confounders, but data that support a role for STH in neurobehavioural outcomes include a study showing that infants and children under 5 years of age with anaemia and STH infection show disturbed social and emotional behaviour. Another study showed that treating school-aged children with antiparasitic drugs and iron supplementation improved attention, memory and processing speed 39 .

GLOBAL RESEARCH PRIORITIES
Prevention of infections that affect the nervous system is the highest research priority, as complete prevention of infection removes all risk of nervous system sequelae. However, treatment of nervous system sequelae and rehabilitation of individuals with nervous system morbidity are also important for the millions who currently live with the nervous system effects of infections. Prevention and treatment of infections that affect the nervous system requires the identification of the pathogens responsible, the pathogen reservoirs and the potential points at which the pathogen life cycle can be interrupted. Table 1 lists specific pathogens, their known nervous system manifestations, and current knowledge gaps regarding incidence and long-term sequelae for each pathogen. Table 2 outlines whether specific interventions (vaccines, control of zoonotic reservoirs or vector populations, and treatment) are available for each pathogen. Table 3 provides a summary of global research priorities for infections that affect the nervous system. These research priorities are discussed in more detail below.

Diagnosis
Improved diagnosis lies at the heart of all research priorities for infections that affect the nervous system because all other research areas depend on accurate infection diagnosis. Improved diagnosis requires better tests to detect infection, better clinical diagnostic algorithms to detect infection and better tools to assess the nervous system effects of infection, including cognitive and mental health sequelae.
Affordable, easy-to-use, rapid diagnostic assays -preferably pointof-care -that can identify infections affecting the nervous system are a high priority. This includes diagnostic tests for infections that directly infect nerve cells and those that do not (for example, malaria and STH). For infections that affect the CNS, field diagnoses are needed to identify when the infection has entered the CNS. Serological assays are available to detect schistosomiasis 36 and cysticercosis, but these tests are not specific for CNS infection, and the blood-brain barrier may prevent the detection of antigens in serum 40 . In the case of bacterial, fungal or viral CNS infections, although lumbar puncture to obtain cerebrospinal fluid is a routine procedure at many centres in LMICs, most lack the capacity for standard bacterial, fungal or viral cultures, let alone more sophisticated testing such as PCR, which is essential for detecting many viral infections. Even in high-income countries with advanced molecular diagnostics, an aetiological agent is identified in less than half of individuals with encephalitis. Because many cases of idiopathic encephalitis are probably caused by viruses still to be characterized, the development of metagenomic and high-throughput screening techniques for viral detection is a research priority, with the goal of eventually developing low-cost diagnostic point-of-care assays for the pathogens identified. For certain CNS infections, notably neurocysticercosis, neuroschistosomiasis, CNS tuberculosis and CNS toxoplasma infection, neuroimaging with CT scans or MRI is needed to make a diagnosis. Although availability of neuroimaging is becoming more widespread, many facilities in LMICs still lack these costly imaging modalities, underscoring the need for research on simple, accurate, low-cost, point-of-care diagnostic tests for detecting infections that affect the nervous system.

S183
One example of how improved diagnostic tools can have an impact is a study 41 from India in which a simple diagnostic algorithm and basic treatment for neonatal sepsis, all performed by village health workers, led to a 63% reduction in neonatal mortality among preterm infants 41 . Simple algorithms for other infections that affect the nervous system, coupled with the ability to provide effective therapy following diagnosis, or appropriate referral for screening algorithms, have the potential to substantially reduce morbidity and possibly mortality from these infections. Improved cross-cultural measurements of neurodevelopment and mental health are a key research priority, and reviewed in the article in this collection on child neurodevelopment (see page S155).

Epidemiology and primary prevention
The lack of affordable, non-invasive, rapid diagnostics for infection and nervous system effects of infection limits our ability to quantify the burden of infection-related nervous system disability ( Table 1). Well-designed studies of disease epidemiology are also required for the accurate measurement of disease incidence, and of the type and duration of nervous system sequelae of infection.
A challenge to the estimation of infection-related nervous system disease is that the symptoms these infections result in, such as epilepsy, hemiparesis or cognitive impairment, are included as 'chronic diseases' in global burden estimates. Careful epidemiological assessment could lead to the more accurate attribution of a portion of 'chronic disease' to its infectious component. For example, the Global Burden of Disease Study 2010 (ref. 42) attributed some of the disability-adjusted life years of epilepsy to neurocysticercosis. This infection is also associated with stroke 43 , which was ranked third in terms of disability-adjusted life years in 2010 (ref. 42), but none of this burden was attributed to neurocysticercosis owing to lack of data. Thus the true burden of nervous system disease owing to neurocysticercosis was probably significantly underestimated.
The most cost-effective method of preventing infection is immunization, discussed in the section on vaccines below. For infections for which there is no immunization, or for which immunization is not highly successful, research is required on sustainable preventive methods. For vector-borne illness, for example, insecticide-based interventions such as insecticide-treated bed nets have reduced malaria incidence and mortality in many areas 44 . But increasing pyrethroid resistance 45 highlights the need for ongoing research even for interventions with documented past success.

Pathogenesis
Disease pathogenesis may be the most neglected research focus of infection-related nervous system disease in LMICs. Although some studies on the pathogenesis of infection-related nervous system disease in individuals in high-income countries are available 46,47 , far fewer studies of pathogenesis have been conducted in individuals from LMICs. Even in high-income countries, studies of infection pathogenesis often use animal models, which may incompletely recapitulate the host response in humans. The host immune response probably contributes to both defence against invading pathogens and subsequent damage to the nervous system 48 , but the type and role of specific cells in the immune response at different infection stages are poorly described. Similarly, it is often unclear which antigens or components of the infecting organism confer neurovirulence. The roles of innate immunity, the microbiome, and co-infection with endemic pathogens, including HIV, in contributing to infection-related nervous system disease are also poorly understood (Box 1). Without an understanding of the pathogenesis of infection-related nervous system disease, it is difficult to rationally plan for adjunctive interventions to prevent or reduce nervous system injury. Although adjunctive interventions have been elusive, those proven successful (for example, steroid treatment in tuberculous meningitis 49 ) have made a major difference in improving neurocognitive and behavioural outcomes. An understanding of the development and types of protective immune responses to antigens or antigenic variants of a pathogen is also fundamental to the development of vaccines, which are, in most cases, the most cost-effective method of preventing infection.

Vaccine development
Vaccines are available to prevent the neurological complications of measles, mumps, rubella, poliomyelitis and varicella virus as well as H. influenzae type b, S. pneumoniae and N. meningiditis. Effective vaccines are also available for rabies and Japanese encephalitis. Together, these vaccines have saved millions of lives as well as prevented long-term nervous system complications in millions of children and adults.
Research priorities for vaccine development include the utilization of disease immunology, epidemiology and pathogenesis studies to develop safe, effective vaccines, and the performance of phase I, II and III trials to determine vaccine efficacy and safety in humans. In P. falciparum malaria, for example, knowledge of antibody and T-cell immune responses to circumsporozoite protein (CSP) 50,51 led to phase I and II trials of the CSP-based RTS,S vaccine 52 . These successful trials led to the recently completed phase III trials of RTS,S 53 . This constituted a major advance in the vaccine field because they established RTS,S as the first successful vaccine in humans against a parasite. However, the relatively modest efficacy (30-50%) of RTS,S was not surprising in light of the known complexity of the human immune response to P. falciparum in endemic populations. Hence, work continues on the development of more effective vaccines. Understanding the human immune response to P. falciparum infection will be key to the development of vaccines with improved efficacy and safety.

Treatment
Treatment with antimicrobials is designed to clear infection or reduce infectious load, decrease disease severity, and ideally to provide a degree of secondary prevention against the nervous system effects of the infection. For viral infections, with the exception of HSV and HIV, there is often no specific treatment. Even for HSV encephalitis, standard treatment (intravenous acyclovir) is unavailable in many parts of the world. Cost effectiveness and stakeholder analyses could be useful in influencing policymakers to increase availability of antiviral

Priority area
Research needed Diagnosis • Rapid, accurate, low-cost, point-of-care diagnostic tests for infections that affect the nervous system • Clinical diagnostic algorithms for infections that affect the nervous system • Improved testing for detection of infection-related nervous-system disabilities Epidemiology • Accurate incidence and prevalence estimates of common infections that affect the nervous system • Accurate identification and frequency estimates of nervous-system manifestations and sequelae • Identification of potentially modifiable risk factors specific to infections that affect the nervous system Pathogenesis • Identification of host response pathways that lead to nervous-system deficits or to clinical immunity • Identification of pathogen factors that lead to nervous-system deficits or to clinical immunity • Assessment of risks and interactions of co-infections and co-morbidity

Operations and implementation
• Optimal methods to implement or operationalize interventions with known efficacy treatment. For HIV-associated opportunistic infections such as cryptococcal meningitis, development of therapies that do not rely on intravenous administration is a priority because capacity for intravenous medication is limited in many LMICs, particularly in rural areas. There is also a need for improved access to new assays that detect antiretroviral therapy resistance (for example, the oligonucleotide ligase assay), as these can guide cART treatment. For many parasitic infections, antiparasitic medications are available, but their efficacy in reducing the neurological or cognitive sequelae remains uncertain. For this reason, as noted in the pathogenesis section, development of adjunctive therapies that target prevention or reduction of nervous-system injury is an important research priority. Development of low-cost, low-toxicity antimicrobials that work against drug-resistant pathogens is a research priority for several infections, including tuberculous meningitis and neonatal sepsis caused by multiresistant Gram-negative infections. Qualitative studies to better understand medical non-compliance, and to develop innovative solutions to reduce non-compliance through newer technologies, such as mobile devices that support medical and public-health practice, are also needed.
Finally, there is a need for multicentre clinical research trials with sufficient sample sizes to provide definitive answers on the efficacy of specific interventions. For example, where smaller trials had failed, the Cryptococcal Optimal Anti-Retroviral Timing (COAT) trial conducted in Uganda and South Africa successfully determined that deferred initiation of anti-retroviral therapy in individuals with HIV until 5 weeks after treatment of cryptococcal meningitis improved survival 54 . This study finding is likely to change international guidelines.

Physical, occupational and cognitive rehabilitation
Whereas physical, occupational and cognitive rehabilitation for individuals with sequelae of CNS infections are routine in developed countries, such interventions are limited in LMICs owing to a lack of trained personnel and prohibitive costs. Thus, research on how to build capacity for rehabilitation and how to support it in the context of LMICs is required. Rehabilitation for cognitive impairment can be successfully implemented in the community using locally available resources or in a tertiary institution using advanced methods. In the community, home stimulation, parenting education and support, and provision of financial support or nutritional support for children enrolled in early child development centres have shown some benefit in improving children's cognition 55,56 . Interventions that target both the carer and the child are more effective than those that include either one 56 . In tertiary centres, computer-based cognitive training programmes have proven effective in improving cognition in African children surviving CNS infections 57,58 . These cognitive training programmes can target specific disabilities; however, they are in their early stages and more research is required to determine the most cost-effective implementable and sustainable programmes for LMICs.

Operations and implementation research
Operationalization and implementation of known effective interventions is another research priority area. Vaccines for S. pneumoniae and N. meningiditis are highly effective and have been implemented in some LMICs, but these vaccine-preventable infections continue to affect more than 1 million people each year. Thus, in addition to increased investment in the basic science of vaccines, a major research priority is the assessment of methods to support and implement widespread vaccination in LMICs.
Another example is implementation research related to effective prenatal, perinatal and neonatal care, which would decrease neonatal sepsis. Assessment of effective methods for non-physician health workers to provide medical and preventive care to mothers and newborns in LMICs are needed 59 . The recent increase in neonatal infection with multi-drug-resistant Gram-negative organisms in LMICs 24,25 makes prevention of neonatal sepsis an even greater research priority. Rapid diagnostics and treatment interventions that are successful in field trials also require implementation research for successful widescale adoption and appropriate use.

Capacity building
Capacity building within LMICs is key to the successful reduction or elimination of nervous system complications of infection. The Review on research capacity building in this collection addresses this topic in depth (see page S207). Research priorities specifically in the area of infection-related nervous system morbidity include an increase in the number of clinicians and researchers in infectious disease, neuroscience, neurology and mental health 60 , and the dedication of a portion of LMICs health budgets to infectious disease and mental and neurological health 61 . Research training grants and collaborative research between partners in LMICs and high-income countries specifically in the area of infection-related neurocognitive impairment can also help to build human resource capacity. Physical infrastructure is another priority; without space for laboratories, diagnostic equipment or research clinics, surveillance cannot be performed and information to guide interventions to reduce the burden of these infections cannot be generated. With improved human resource capacity and infrastructure, the development of effective screening instruments, prevention

Diarrhoeal disease
Diarrhoea is a leading cause of mortality in children living in low-and middle-income countries (LMICs) 62 . Most of these deaths occur in Africa, Southeast Asia and in eastern Mediterranean countries 63 . In 2010, it is estimated that there were 1.731 billion episodes of diarrhoea worldwide of which 36 million progressed to severe diarrhoea and 700,000 episodes resulted in death 64 . Its occurrence in the first 2 years of life is associated with an 8 cm decrease in height and a 10 point drop in IQ by the time children are around 7 to 9 years old 65 . The mechanism by which diarrhoea affects cognition is not clear, but it could be through the effect of diarrhoea on stunting, which in turn predicts future cognition 66 . However, during diarrhoea-free periods in the first 2 years of life, children experience catch-up growth and may return to their original growth trajectories 67 . This highlights the importance of effective interventions for diarrhoea to sustain the child's developmental potential. The high frequency of diarrhoea episodes during this critical developmental stage and the large number of cases makes diarrhoea a major public health concern for child development.

The microbiome
Although it has been clearly demonstrated that pathogenic microbes can cause brain disorders, there is increasing evidence that the microbial population harboured in the human body, termed the human microbiome, can as a whole influence brain activity 68 . Recent clinical studies among healthy subjects suggest that treatment with a probiotic is associated with reduced symptoms of stress and depression 69 . There is also evidence of associations between the microbiome and neurological diseases, such as multiple sclerosis and autism spectrum disorder (ASD) 70 . In a recent study using a mouse model of ASD, treatment with probiotics alleviated some behavioural symptoms of the disorder 71 . The composition of the human microbiome shows marked differences between countries 72 and comparative research conducted in highincome countries and LMICs could lead to a better understanding of the part played by the human microbiome in brain disorders, and possible treatment of these disorders with factors that favourably alter the microbiome. S185 and treatment, and increased government support to address these infections are more likely to be achieved. Box 2 provides examples of capacity building in Peru and Uganda in the area of infections that affect the nervous system. This was enabled by Fogarty International Center grants for collaborative US-LMICs partnerships in these countries.

CONCLUSIONS
In the past two decades, an increasing body of evidence implicates infection as a cause of substantial nervous system morbidity in high-income countries and LMICs. The burden is particularly high in LMICs, where infections such as HIV (and associated opportunistic infections), HSV, dengue, bacterial and tuberculous meningitis, malaria, neurocysticercosis, STH, schistosomiasis and other infections affect billions of people annually, and cause substantial neurological morbidity in those individuals. The involvement of the nervous system in infections is often first recognized in LMICs where the prevalence of these infections is higher and where new infections often emerge. Research conducted in these countries can contribute to prevention and cure before these infections become globalized. Research is needed in multiple areas to determine the true burden of disease and to develop point-of-care diagnostic assays for diagnosing infection, vaccines and other interventions for preventing infections; to improve our understanding of the pathogenesis of nervous system disease in these infections; to develop better tools for the assessment of neurological, cognitive and mental health impairment; to develop more effective treatments and preventions for nervous system sequelae, to improve the implementation of successful interventions, and to improve rehabilitation for those with long-term neurocognitive or mental health disabilities. Good research studies in these areas, accompanied by equally strong efforts to implement promising technologies and therapies, could substantially decrease the morbidity and mortality of infections affecting the nervous system in LMICs.

Uganda
The Severe Malaria Research Centre in Uganda is an example of how collaborations between local and foreign scientists, with support from the Fogarty International Center, has built a hub for research excellence. Through a National Institutes of Health (NIH) R21 exploratory research grant in 2004, local scientists developed research capacity by involvement in research studies, and grant and manuscript writing. This has since led to further NIH grants (four R01 grants, a U01 grant, a D43 grant and two R34 grants) as well as multiple grants from other agencies. Ugandan and US faculty are principal investigators on these grants. A book on neuropsychology of African children and more than 30 research papers have been published from these projects so far. Ugandan scientists and physicians have obtained faculty positions in Makerere University, Kampala. The infrastructure that has been built includes high-speed Internet connectivity for research offices and faculty members, a laboratory, a data room and a grants management office. With the infrastructure in place, this centre is now providing training for many Ugandan and US students and researchers at all levels of training, from undergraduates to post-doctoral fellows and faculty.

Peru
Through the Fogarty International Center NIH R21 and R01 grants in Peru, a network of neurologists who are engaged in braindisorder research has been developed throughout the country. Trainee alumni of this network now serve as collaborators on emerging research and training activities in both infectious and chronic diseases of the nervous system (such as cerebrovascular diseases). The 2 sites in Lima have been scaled up to 12 hospitals and 2 universities in 3 Peruvian regions. Capacity building of individuals was provided through workshops, hybrid virtual/ in-person certificate courses, as well as medium-and long-term training in Seattle and Peru. An initial mentor-training workshop developed into a growing network of mentors, three of whom have been awarded Clayton-Dedonder Mentorship Fellowships by the Fogarty International Center and have started institutionalization of mentor training programmes at three institutions in Lima. Those who received the R21 and R01 grants in the past are now experienced researchers who are leading the development of research in new areas, such as neurogenetics research and the development of a cerebrovascular diseases research training programme. Research supported by these awards resulted in 26 peer-reviewed publications and book chapters. Programme alumni are becoming leaders in brain research and are mentoring the newest wave of young neurologists and neuroscientists.