Lack of Diversity in Human Neuroscience Research
Although neurosciences have made notable progress in theory and technology, a significant methodological issue persists with the continued reliance on homogeneous sample groups (Burns et al., 2019). This observation is neither new nor recent, as several authors have already raised concerns about the fact that behavioral scientists base their conclusions mostly on samples composed almost exclusively of people from Western, Educated, Industrialized, Rich, and Democratic (WEIRD) societies, even though these groups are very peculiar, representing only 12% of all humans and may be considered outliers (Henrich et al., 2010). The problem is exacerbated by the frequent use of convenience samples, which tend to consist primarily of university students, resulting in a narrower demographic range that fails to represent the broader diversity within society.
The inclusion of more diverse populations is crucial, notably in light of the nature versus nurture debate. It is indeed now widely accepted that both cultural and biological factors contribute to shaping human psychology and behaviors (Chiao and Ambady, 2007). However, the lack of diversity in population samples diminishes the impact of cultural (nurture) factors in the obtained results and makes it difficult to claim whether neurocognitive processes are universal or culture specific (Burns et al., 2019). Beyond cultural differences, the lack of access to diverse populations can hinder the exploration of innovative research questions in the field, which may require targeting specific populations that use, for instance, different metric systems, have unique hierarchical structures, or have experienced a unique historical event.
A critical question thus remains: Why does the pursuit of greater diversity in population samples remain so limited? Part of the answer is undoubtedly tied to the unequal distribution of funding and the number of scientists in specific parts of the world who lack the resources to conduct research in human neuroscience. However, in countries where researchers have access to more financial resources and can conduct research projects abroad, the answer may lie in the operationalization and practical aspects of how to engage with more diverse samples. Indeed, many researchers and financial institutions have deep concerns about the feasibility of conducting field research in psychology and neuroscience. It has been argued that although the issue of focusing on too-narrow samples may be easily solvable in disciplines using methods relying on papers and pens, disciplines such as neuroscience face additional challenges because of the equipment needed. In many countries, some of the most used machines in neuroscience [e.g., Magnetic Resonance Imaging (MRI); near-infrared spectroscopy (NRIS), electroencephalography (EEG), transcranial direct current stimulation (tDCS), and transcranial magnetic stimulation (TMS)] are not available because they are too expensive or, if available, are not adapted for experimental protocols. However, because many methods are now increasingly portable, there is nothing preventing researchers from conducting their projects outside of their laboratory, or even abroad. This would enhance the diversity of population samples, thereby allowing broader generalization of the findings and yielding numerous associated benefits, such as a more equitable distribution of knowledge and the initiation of novel scientific collaborations.
This article aims to provide practical guidelines and solutions for conducting neuroscience projects in diverse and sometimes challenging environments, including strategies to involve more diverse samples and prevent “systematic racism” in neurosciences (Parker and Ricard, 2022). The guidelines are drawn from studies using electroencephalograms (EEGs) to assess group dynamics and trauma in a postgenocidal context, which required accessing sensitive populations that often lived in remote rural areas in Rwanda and Cambodia (Caspar et al., 2023). I will present various challenges and solutions associated with conducting such research projects, including ethical considerations, transporting materials, recruiting participants, and cleaning noise in EEG data, all of which frequently impede the achievement of diverse sampling in neuroscience. Important resources researchers should consult prior to initiating such projects will be introduced, but I will also openly discuss the practical difficulties that arose in my studies to better represent the reality of the process. I will also detail some of the costs requested for such research projects that go beyond the usual equipment or personnel expenses. Some funding agencies, for instance, do not cover certain types of costs (e.g., visa and travel insurance fees, fees charged by local ethics committees, per diems, additional health precautions such as vaccines, etc.), so it is crucial to check these aspects beforehand. Translation costs, notably, can be high, as they are required at various stages (e.g., information and consent forms for participants, local translators, etc.).
Guidelines and Insights to Diversify Population Samples
Ethics approvals from local infrastructures
Researchers in the human sciences must obtain ethics committee approval before initiating a study. These committees safeguard participants’ rights and well-being, ensure informed consent, and prevent harm. However, ethical codes are not universal and can vary depending on local regulations, necessitating approval from local ethics committees. Yet this step is sometimes overlooked by external researchers who may adopt a patronizing attitude, assuming they know better (Schroeder et al., 2021) and request an ethics approval from their own institutions instead of the local ones. Particular attention should be paid to the phenomenon of ethics dumping phenomenon (Schroeder et al., 2021), which refers to the unethical practice of conducting research in low-resource or vulnerable settings where ethical standards are lower or less strictly enforced. This often involves researchers from higher-income countries carrying out studies in lower-income countries in ways that would not be permitted in their own countries due to stricter ethical guidelines. Many countries do have infrastructures for conducting ethics evaluations. A useful resource in this regard is a recent paper by Köhler and colleagues, who interviewed 87 out of 146 national ethics (or bioethics) committees around the world (Köhler et al., 2021) and provides researchers with a list of countries where established ethics committees are operational.
It is important to note that some countries currently lack the infrastructure for ethics committees, which requires researchers to seek permission in their own countries. In such cases, it is the researchers’ responsibility to ensure that their projects adhere to local regulations and respect the local populations and cultural sensitivities. A valuable resource, discussed in the following section, is the “Global Code of Conduct for Research in Resource-Poor Settings” (GCC; TRUST, 2018), developed by the European Commission to crack down the ethics dumping phenomenon.
In practice, the process of obtaining ethics approvals may require fees and take several months. For example, in Cambodia, the platform for the local ethics committee was not operational, and it took ∼8 months to finally obtain an answer. The submission procedure had to be conducted in person in Phnom Penh, which was 9,900 km (∼6,050 miles) away from my current location and required a fee of $400. Additionally, they did not provide advice for studies outside of the medical field and directed me to a local university. In Rwanda, online submissions were authorized with a fee of $1,500. Researchers should also be mindful of the cost of translation here, because information letters and consent forms should of course be provided in the native language of the participants, and possibly the ethics application form itself, depending on the country.
Local collaborations and data transfer agreements
Ethics committees or institutions may request collaboration with a local researcher for project acceptance. While this step may not be mandatory in all countries, it is nonetheless essential as it enhances the relevance and impact of research findings and ensures fair knowledge transfer. In this context, numerous institutions worldwide, including the European Commission (EC) and various publishers, have adopted the TRUST code, already mentioned to help fighting ethics dumping (TRUST, 2018). Available in several languages, this resource provides guidance for all researchers in ensuring their international research is equitable and demonstrating how partnerships between high-income and lower-income countries can benefit both parties.
Insights from local academics and stakeholders are important for designing and implementing international research projects. In Rwanda and Cambodia, however, I was unable to find experts in cognitive and social neuroscience, so I reached out to academics in psychology and sociology. It is worth noting that your academic institution might already have agreements with foreign universities in various disciplines, which could ease the search for a collaborator. Some embassies, with their research and development sections, are also familiar with local universities and researchers and are a valuable resource for assistance in this process. In practice, the absence of local researchers in a field may create challenges with authorship, as journals often require proof of significant contributions from coauthors. Listing someone without a major role is considered unethical. Although finding support in such disciplines may be difficult, assistance with local regulations, funding, and participant access is often still possible. To ensure fairness, authorship and collaboration should be discussed early on, involving local collaborators extensively.
When collaborating internationally, researchers must be cautious about the accessibility, transfer, and exchange of collected (personal) data. Within the European Union (EU), the General Data Protection Regulation (GDPR) sets strict rules for handling data, including its transfer outside the European Economic Area (EEA). Consequently, researchers must understand the regulations of their own country, the regulations set by their funding sources, and the laws of the country where the research is being conducted. While not every country has enacted specific data transfer laws, the ethical obligation to obtain informed consent from participants and to safeguard data privacy is widely recognized as a part of conducting ethical research. Open Science practices have facilitated the sharing of (pseudo-)anonymized data. However, funders often require assurances that personal data handling complies with both their standards and local regulations. For example, data intended for transfer from a non-EU country to the EU must be justified to show compliance with the data's country of origin. In the absence of specific laws on data transfer, it is advisable to establish a data transfer agreement between the involved parties, something research institutions can initiate.
Practical aspects: material transport and testing sites
In the research projects conducted in Rwanda and Cambodia, we needed to transport our equipment from Belgium to these countries and back. While we could handle paperwork directly in the target country with printing stores, the neuroscience equipment was significantly more delicate and heavier. Typically, our material weighs between 20 and 50 kg, which includes one or several electroencephalogram (EEG) machines, two to six portable computers, electrode sets, conductive gel, batteries, chargers, and other smaller items. I would first advise securing international insurance for your materials through your institution and before addressing the challenge of transportation.
A straightforward option is to purchase extra luggage allowances from airline companies. However, there is a widespread awareness of the instability of this process with thousands of bags displaced or lost by airlines companies annually. To avoid joining these statistics and to minimize the risk of damage during transit, I recommend diplomatic bag services. This involves requesting that the department of external affairs in the home country sends the materials through diplomatic channels to the embassy in the destination country. Typically, delivery takes ∼10 d. As outlined by other researchers (Jasińska and Guei, 2018), local authorities reserve the right to inspect checked baggage at the airport, which may damage fragile equipment or cables during inspections. The diplomatic bag services employ a special lock to secure the contents, which can only be opened by embassy staff, thus drastically enhancing the safety and security of the transport. In addition, it has the advantage of protecting any data inside the box (e.g., consent forms; hard drive; interviews) and prevent uncontrolled and unplanned individuals from accessing such data, a critical element for ethical consideration.
Using diplomatic bag services typically involves a charge of approximately €500 for transporting 30 kg of materials one-way. Additionally, this system is only possible if your country has an embassy in the destination country. For instance, in Cambodia, Belgium has only a consulate in Phnom Penh, not an embassy equipped to handle diplomatic bags. Therefore, we had to send the bags to the nearest Belgian embassy in Bangkok, Thailand, and then rent a car to transport the materials across the Thai-Cambodian border, a 3–4 h drive. Another complication can be batteries; the diplomatic bag service sometimes restricts us to including no more than two batteries per box. We must therefore carry any additional batteries in our carry-on luggage, ensure they comply with aviation regulations, and carry with us the appropriate documentation.
Finding a suitable location for testing participants is challenging but crucial for obtaining reliable (EEG) data (see below, Brain measurement in the field). While universities may have available rooms, they are rarely dedicated to research. At the University of Rwanda and the National University of Battambang, for example, no experimental rooms were available. We often used colleagues’ offices or empty classrooms, but availability varied during the academic year. In rural areas, this becomes even harder due to the lack of electrical outlets. In Rwanda, we tested in churches or bars, and in Cambodia, we rented small shops. Some researchers recommend using customized tents with opaque, waterproof roofs and walls as portable experimental rooms (Jasińska and Guei, 2018).
Brain measurement in the field
Portable neuroimaging and electrophysiological methods provide new advances to the study of brain function in remote locations with previously inaccessible populations. I rely on EEG because it can be easily transported, is riskless for participants, does not take too much space, and is not too heavy to transport. Other portable technologies are also available, notably fNIRS (Burns et al., 2019; Jasińska and Guei, 2018).
EEG signals can be affected by a great variety of artifacts, including power line interference at 50 or 60 Hz. Shielded rooms, such as Faraday cages, are unlikely to be found in the field. In university labs in Europe, most of the electrical circuits are correctly grounded for security reasons and the power line noise is stationary. Therefore, even if the testing takes place outside of a Faraday cage, a simple spectrum interpolation at 50 Hz and its harmonics (i.e., 100, 150 Hz) is a reliable technique to clean the data. In rural Rwanda and in Cambodia, however, the locations where we tested had very rudimentary electrical circuits, which led to multiple nonstationary power line noises. To get rid of these, the ZapLine toolbox developed by de Cheveigné (2020) can be useful, as it can remove power line artifacts from multichannel data by applying spectral and spatial filtering. Another effective technique is to unplug laptops and avoid using external screens that require plugging in, as they can transfer noise to electrodes. Using external batteries can ensure computers last all day, but it is essential to check airline and diplomatic baggage regulations regarding battery allowances. Alternatively, some researchers have recommended using portable solar generators to power small- to medium-sized electronics or using a diesel generator (Jasińska and Guei, 2018).
Another important element is the size of the EEG caps. People from North America, Australia, and Europe are in average taller than the populations located elsewhere, which is generally associated with bigger head circumference under normal nutritional and health conditions. In rural areas in Rwanda and in Cambodia, the prevalence of stunting in children is elevated, which may alter body development and head size. This may result in floating headcaps, which can alter the EEG signal and source reconstruction.
It has long been acknowledged that there is systematic exclusion, sometimes even structural racism (Parker and Ricard, 2022), of people with certain skin colors or hair types from neuroscience research. This systematic exclusion is partly due to technical and methodological limitations in neurotechnology. However, it is also the result of researchers directly excluding participants with specific hairstyles without attempting inclusion, which perpetuates common practices in the field. This has consequences not only for the generalizability of the research results but also contributes to reported distrust in research among these excluded populations. With several systems and a bit of expertise, it is entirely possible to include these individuals in research projects instead of excluding them. For example, when we conduct tests in remote areas of Central Africa, the majority of the women have short hair, which is particularly suitable for EEG recordings. In urban areas or among people of African descent in Europe or in the United States, more elaborate hairstyles or braided hair are common. However, even for these individuals, it is often possible to acquire very good EEG signals using a classic 32-electrode EEG system. Therefore, it is recommended that one first practices with one's own system before deciding the exclusion of participants. Another possibility for a better inclusion is the preparation of participants before recordings are made, instead of a direct exclusion. For instance, the Inclusive EEG Handbook (https://www.inclusiveneuro.com/) presents practical information on how to do brain sensing on textured hair.
Temperatures, ground component, and humidity can also affect the equipment, the data acquisition, and the participants’ comfort. In environments where humidity is high, typically exceeding 60%, electronics are at a higher risk of corrosion. In some countries, during the rainy season, humidity can increase up to 95%, thus impacting the material. Dust can cause additional damage to the equipment. High temperatures and humidity can also cause sweat artifacts in EEG recordings, and this type of noise cannot reliably be cleaned. While a solution is to avoid testing during the rainy season and ensuring proper maintenance of the equipment, portable air conditioning units can also be used, if the power supply allows it (Jasińska and Guei, 2018).
In addition to addressing practical aspects of the EEG equipment, we also dedicate a significant amount of time to preparing participants and explaining the equipment and its purpose. Upon our arrival in remote villages in Rwanda and Cambodia, the presence of foreigners, computers, and unfamiliar technology is often completely new to the villagers. Some villagers may be intimidated by the equipment, fearing that it could read their minds or affect their physical or psychological health. We may compare the EEG to checking for a fever: just as you would place your hand on someone's forehead to feel if they’re hot, the EEG “touches” their head to “listen” to what their brain is doing, without injecting anything or reading their thoughts. We have also observed that many participants are surprised when events repeat multiple times during the computer task—a common practice in neuroscience research. Therefore, we systematically explain that they will encounter repeated images or sounds because the brain “whispers,” or “does not speak loud,” and we need to present the information multiple times to collect reliable data.
Given all the abovementioned complexities, when preregistering studies, it is recommended to mention the possibility of testing additional participants to compensate for data loss due to challenging experimental conditions.
Final Remarks and Conclusion
The present article intends to address the challenges associated with concrete operationalization to enhance diversity in the populations studied in neuroscience. Beyond the direct engagement of researchers, several recommendations can also be made for publishers and funding agencies. Firstly, echoing previous research, each scientific publication should include a Constraints of Generalizability (CoG) section to provide a clearer understanding of the findings’ generalizability, which is crucial for interpretation and replication. Secondly, funding agencies possess the leverage to support or prioritize projects that incorporate cultural diversity and can offer training and resources to facilitate inclusive practices among researchers. An example of this approach is the NIHR-Include project, which offers guidelines on enhancing the participation of underrepresented groups in research and outlines critical considerations for researchers, funders, and reviewers. While this project is primarily focused on clinical research within high-income countries and does not encompass international research in low-income regions, its principles can guide researchers aiming to diversify their sample populations domestically.
By adopting these strategies, researchers, funding bodies, and publishers can collectively ensure that research samples more accurately represent the vast diversity of the global population. This, in turn, enhances the generalizability and relevance of research findings. Promoting inclusivity in research will not only facilitate broader knowledge exchange across the world but is also essential for scientific progress. Inclusive research practices provide deeper insights into neurological variability and are crucial for understanding the intricate interplay between genetic and environmental factors.
Footnotes
This work was supported by a European Research Council Starting Grant DISOBEY (grant number: 101075690) and by a Bijzonder Onderzoeksfonds UGent Starting Grant from Ghent University (grant number: BOF/STA/202109/025) granted to E.A.C. The author used ChatGPT (OpenAI) only to check for the presence of typos or possible grammatical mistakes in the manuscript, as English is not her mother tongue. After using this tool/service, the author reviewed and edited the content as needed and takes full responsibility for the content of the publication.
Views and opinions expressed are however those of the author only and do not necessarily reflect those of the European Union or the European Research Council Executive Agency. Neither the European Union nor the granting authority can be held responsible for them.
The author declares no competing financial interests.
- Correspondence should be addressed to Emilie A. Caspar at emilie.caspar{at}ugent.be.
