Two decades of neuroscience publication trends in Africa

Africa bears 25% of the global disease burden despite comprising 15% of the world’s population, and hosts exceptional biodiversity relevant to understanding neurological disorders. Yet neuroscience research output from the continent has lagged due to low funding, inadequate infrastructure, relatively few active scientists, and heavy administrative and teaching loads. Prior bibliometric estimates using database search terms could not reliably distinguish African-led studies conducted in African laboratories from work led elsewhere that included African affiliations. For example, substantial proportions of papers affiliated to African countries are led by non-African researchers, and many African-led papers appear in local journals with limited international visibility. To address this gap, the authors systematically identified and manually curated neuroscience publications affiliated with African institutions from 1996–2017 to quantify truly African-led work carried out within Africa. They extracted metrics on output by country and sub-discipline, journals and citations, funding and international collaborations, experimental models and techniques, and compared these patterns with samples from the UK, USA, Australia, Japan and Brazil. The overarching research questions were: how many African-led neuroscience publications exist by country over time, how sub-disciplines are represented, how collaboration and funding patterns shape visibility, and what techniques and models are employed continent-wide.

The paper situates its analysis within prior observations that African neuroscience output has grown in some countries (e.g., South Africa, Egypt, Nigeria) but often involves high proportions of non-African-based authors. Previous estimates in specific settings (e.g., Burkina Faso) suggested up to 80% of health research including African authors was not African-led. Database mining approaches using keywords (e.g., “neuroscience” and country names) tend to overcount by including non-African-led work; for Nigeria, PubMed mining identified 127 neuroscience papers (1996–2017), but manual curation found 54% led by non-African researchers. Many remaining African-led papers were published in Africa-based journals with relatively low citation rates and limited indexing. Prior commentary also highlights limited use of resource-intensive, modern techniques (e.g., fluorescence microscopy, neuroimaging) and genetically modified models across parts of Africa, reflecting infrastructural constraints. These gaps motivated a continent-wide, manually curated assessment of African-led neuroscience to produce more accurate, comparative metrics.

Study design: Bibliometric and metadata analysis of neuroscience-related primary research publications affiliated with African institutions from January 1996 to December 2017, with manual curation to ascertain African leadership and that work was performed in Africa. Comparative samples were drawn from five non-African countries (UK, USA, Australia, Japan, Brazil). Data sources and search strategy: Publications were retrieved from PubMed using combined neuroscience-related terms (e.g., “neuroscience”, “neuro”, “neuron”, “spinal cord”) and country names. Searches included author affiliation fields and, where applicable, full text. Review articles and irrelevant items were excluded; duplicates were removed. Inclusion/exclusion and manual curation: Each PubMed-listed item was assessed by hand to determine if it was (i) African-led (based on lead/senior author affiliations) and (ii) presented clear evidence that the research was carried out in Africa. Items failing these criteria (e.g., work led entirely from outside Africa or conducted elsewhere despite African affiliations) were excluded. Of 12,326 PubMed-indexed items with African affiliations, manual curation eliminated approximately 58%, yielding 5,219 African-led, Africa-conducted publications for analysis. Comparative dataset: Using the same criteria, a set of randomly sampled publications from the UK, USA, Australia, Japan, and Brazil over the same years was assembled to extract comparable metrics. Data extraction and variables: For African-led papers, the team extracted author affiliations, country, year, sub-discipline classification (nine broad neuroscience topics), publishing journal and Clarivate Analytics impact factor (IF), citation counts, funding acknowledgements (presence/absence; domestic vs international; specific funders), international co-authorship patterns (within and beyond Africa; by region/country), model systems (including transgenic use), and laboratory techniques. Techniques were classified into Type 1 (minimal infrastructure, e.g., classical histology, chromatography, behavioural assays) and Type 2 (resource-intensive, e.g., fluorescence microscopy, molecular biology, cell culture, neuroimaging). Use of endemic medicinal plants was also coded. Quality control: Curation was performed in duplicate by different team members; discrepancies were resolved by the first author after discussion. Data management used MS Excel and R; analyses employed Igor Pro and R. A custom database of international flight routes (OpenFlights.org) was used to compare co-authorship networks with air connectivity. Data availability: All curated data and code resources are publicly available (GitHub: https://github.com/podalakol/AfricanNetworking2020).

• Output volume and distribution (1996–2017): After manual curation, Africa produced 5,219 African-led neuroscience publications. Five countries accounted for over three-quarters: Egypt (1,478; 28%), South Africa (1,181; 23%), Nigeria (566; 11%), Morocco (409; 8%), Tunisia (388; 7%). East African contributors included Kenya (131; ~2–3%), Ethiopia (119; ~2–3%), Tanzania (103; ~2–3%). Cameroon (81), Malawi (71), Algeria and Senegal (76–80), Uganda (69), and Ghana (60) each contributed 1–3%. More than half of African countries produced fewer than 10 papers. Annual output increased exponentially across regions over the period.
• Sub-disciplinary representation: Topic distributions across African neuroscience were broadly stable over time and similar across countries, though some areas (e.g., sensory systems) were relatively less prevalent compared with outside-Africa samples.
• Citations and journal impact: African-led papers had few entries in the highest citation/IF brackets relative to non-African regions. Citations correlated positively with journal IF. Very few African publications fell into the top citation bracket (≥95 citations) and high IF brackets.
• International co-authorship: African-led papers with international co-authors (within Africa or beyond) had significantly higher citations and appeared in higher IF journals than solo-African papers (Wilcoxon rank-sum tests: T ≈ 1.74×10^-15 for citations; T ≈ 1×10^-12 for IF; n = 45 countries). Co-authorship networks showed sparse long-distance intra-African links; collaborations more commonly bridged to Europe and North America. Within Africa, Southern Africa was the most frequent intra-continental partner; North African papers often included co-authors from Europe, North America, and the Middle East. Geographic and cultural/historical ties, and logistical factors (e.g., flight connectivity), mirrored observed collaboration patterns.
• Funding patterns: Among 265 top African papers (IF ≥ 5), 35% (n = 93) declared no funding. Of those declaring funding (n = 172), most listed international rather than domestic sources. Southern Africa was the only region where domestic funding mentions exceeded international (domestic 73% vs international 48% mentions among papers declaring any funding). In East Africa, only 8% declared domestic funding while 97% declared international support. Across all African top papers that declared international funding (n = 123), the USA was cited most frequently (36% of international-funding mentions), followed by the UK (25%), France (9%), and Switzerland (6%). Top international funders acknowledged included the NIH (34% of international mentions), Wellcome Trust (20%), MRC (8%), and WHO (5%). By comparison, in non-African samples, 92–100% of papers declared domestic funding, with additional international support varying by country.
• Techniques and models: Use of resource-intensive Type 2 techniques was low across Africa. No country exceeded 25% Type 2 usage except The Gambia (36%, linked to an MRC unit). Examples: Egypt 25%; Tunisia 10%; Algeria 6%; many countries <5% or 0%. In contrast, Japan, UK, and USA published ~75% of papers using Type 2 techniques; Australia 54%; Brazil 33%. Use of genetically modified model systems was almost entirely absent in Africa: no African country exceeded 1% of publications using transgenic models; most used none. Non-African comparators used genetically tractable models widely. Research on endemic medicinal plants was prominent in several West African countries: Cameroon (30% of papers), Nigeria (18%), Ghana (18%), whereas Egypt (5%) and South Africa (1%) focused more on other areas.
• Overall trajectory: African neuroscience output and workforce are growing, but visibility and technology adoption remain uneven across regions, with West Africa showing particularly low representation in high-IF venues.

The study demonstrates a clear and accelerating rise in African-led neuroscience publications since 1996, yet documents strong concentration of output in a handful of countries and persistent gaps in visibility and technological capacity. International co-authorship is associated with higher citations and publication in higher-IF journals, underscoring the benefits of global networking. However, intra-African collaboration remains sparse, suggesting that strengthening continental research networks could further enhance impact and resilience. Funding patterns reveal a heavy reliance on international support, with Southern Africa (notably South Africa) standing out for higher domestic funding mentions, likely reflecting national R&D policy and increased GERD over time. Broadly, sustainable growth in African neuroscience will require expanded domestic funding streams (governmental and philanthropic) alongside continued international partnerships. The limited access to and use of resource-intensive techniques, and the near absence of genetically tractable model systems, likely constrain both the scope and visibility of African neuroscience. Given cost-effectiveness and transformative impact of models like Drosophila, zebrafish, and C. elegans, enabling their adoption—through investment, training, open hardware, and appropriate biosafety regulations—could markedly improve research capacity. Similarly, ensuring equitable access to shared core facilities and promoting publishing in indexed journals (with fee waivers where needed) could bolster international visibility. The findings address the initial research questions by quantifying African-led outputs, mapping sub-disciplinary representation, showing the positive influence of international co-authorship on impact metrics, documenting predominant reliance on international funding, and revealing limited use of advanced methods and transgenic models. Together, these insights frame targeted strategies to enhance Africa’s neuroscience landscape.

This study provides a manually curated, continent-wide assessment of African-led neuroscience research from 1996–2017, quantifying outputs by country and topic, collaboration and funding patterns, and usage of techniques and models, with comparative context from five non-African countries. Key contributions include: (i) a robust count of African-led publications (5,219) and their geographic concentration; (ii) evidence that international co-authorship correlates with higher citations and IF; (iii) documentation of heavy reliance on international funding; and (iv) identification of substantial gaps in advanced technique adoption and genetically tractable model use. Future directions should prioritize: expanding domestic funding (governmental and philanthropic); facilitating intra-African collaboration and mobility; improving access to modern equipment, training, and open hardware; establishing regulatory frameworks for genetically modified organisms; and encouraging publication in indexed journals with accessible fee structures. These steps could increase both the volume and visibility of African neuroscience and may generalize to other scientific disciplines on the continent.

• Database scope: Reliance on PubMed-indexed records excludes non-indexed African journals, potentially underrepresenting local outputs.
• Manual curation: Although curated in duplicate with consensus resolution, human judgment may introduce classification errors in determining leadership and locus of research.
• Time frame and sampling: The analysis is limited to 1996–2017; patterns may have shifted since. Funding analyses focused on a subset of top papers (IF ≥ 5), which may not represent all outputs.
• Metric constraints: Citation counts and journal impact factors are imperfect proxies for research quality or societal impact.
• Comparative sample: Non-African comparator datasets were sampled and filtered under similar criteria but may not perfectly match African datasets in size or composition due to exclusion criteria.