Draft — Global Health Emerging Scholars Fellowship

MOTIVATION LETTER

Substance use disorders are rising across sub-Saharan Africa, yet the computational tools to model their neurobiological drivers remain almost entirely developed in and for high-income settings. My research directly addresses this gap. I am Eniola Ayodele Olutogun, an independent computational researcher and licensed pharmacist from Nigeria, and I am applying to the Global Health Emerging Scholars Fellowship to advance my Conjunctive Consolidation Threshold model of reward-memory encoding in addiction.

The CCT model is a tripartite pharmacological framework I developed and tested entirely as an independent researcher. It couples dopaminergic reward-prediction-error signals, NMDAR-dependent long-term potentiation, and affective contrast into a single ordinary differential equation system. I calibrated the model using Bayesian MCMC with literature-elicited priors drawn from a systematic screen of 1,847 records. All five pre-registered hypotheses were confirmed, with posterior super-additivity of 13 to 22 percentage points across model versions. Three sole-authored preprints are deposited on OSF and Zenodo, and a co-authored paper is under review at Alcohol.

This fellowship would allow me to spend three to six months at a U.S. host institution, working with a mentor to fit the CCT model to real behavioral data from human or animal studies. My neurocascade simulation engine, which couples pharmacokinetics to receptor-binding to Wilson-Cowan circuit dynamics, is ready to serve as the translational bridge. The engine passes 62 of 62 tests and is calibrated for mu-opioid, D2 dopamine, and GABA-A receptor systems. A U.S. visit would provide access to experimental collaborators and datasets that are not available in Nigeria.

My academic record includes a B.Pharm from the University of Ibadan with a German-equivalent grade of 1.9, and I am enrolled in the M.Sc. Digital Health programme at the Hasso Plattner Institute in Potsdam, Germany, starting winter 2026. My independent research output includes pre-registered studies on hERG cardiotoxicity topology, drug-resistance mutation prediction using protein-language models, and Koopman-operator methods for drug transport in tumor tissue. I have endorsements from Kent Berridge at Michigan, Samuel Gershman at Harvard, Nathaniel Daw at Princeton, and Marcelo Mattar at NYU.

The Global Health Emerging Scholars Fellowship is the right programme for this work because it explicitly supports early-career researchers from LMICs, values rigorous computational approaches, and requires a U.S. visit that would directly enable the collaboration my models need. I am ready to commit to that visit and to return to Nigeria with a clinically actionable computational framework for addiction neuroscience.

RESEARCH STATEMENT

My research programme addresses a specific global health problem: the lack of computational models that can predict, at the individual level, whether a given pharmacological intervention will prevent the consolidation of reward memories in substance use disorders. The standard approach in addiction neuroscience relies on qualitative circuit diagrams and single-pathway hypotheses. I build quantitative, testable models that integrate multiple neurotransmitter systems and produce falsifiable predictions.

The CCT model is the core of this programme. It posits that reward-memory consolidation requires the conjunctive activation of three axes: a dopaminergic reward-prediction-error signal above a threshold, NMDAR-dependent long-term potentiation above a separate threshold, and a positive affective contrast signal. The model is implemented as a system of coupled ODEs solved with RK45 and calibrated with Bayesian MCMC using PyMC's DEMetropolisZ sampler. The 14 free parameters were assigned priors from a systematic literature screen of 1,847 records. All five pre-registered hypotheses were confirmed. The model predicts that partial blockade of any single axis is insufficient; only simultaneous reduction of all three below their thresholds prevents consolidation. This has direct implications for combination pharmacotherapy in addiction.

The neurocascade simulation engine extends this work by modeling the full chain from drug pharmacokinetics to receptor binding to circuit dynamics to behavioral readout. It is implemented in Python with ODE layers for each stage and calibrated with Bayesian methods for three receptor systems. All 62 unit tests pass. The circuit-layer parameters are currently illustrative, pending fits to real behavioral data. That is the next step, and it requires access to experimental collaborators and datasets.

My other research lines demonstrate the same methodological approach. The hERG cardiotoxicity study was a pre-registered, powered replication that tested whether bipartite persistent homology could predict cardiotoxicity from protein-ligand interface geometry. It found that topological features do not beat a plain descriptor baseline, settling a comparison the literature had never actually run. The interface-topology-for-resistance study applied the same methods to drug-resistance prediction and found they carry almost no signal, ruling out interface geometry as the driver. The TOPOLOGIX project then built a sequence-based predictor using ESM-2 protein-language-model delta-embeddings and Morgan fingerprints, achieving AUROC 0.804 on the Platinum benchmark and covering 100% of mutations versus approximately 18% for structure-limited tools.

The fellowship would fund a three-to-six-month visit to a U.S. host institution where I would fit the CCT model and neurocascade engine to behavioral data from animal models of addiction. The expected outcome is a validated, open-source computational framework that predicts the efficacy of combination pharmacotherapies for substance use disorders. This framework would be directly applicable to the Nigerian context, where opioid and methamphetamine use are rising and where computational pharmacology expertise is scarce.

SHORT ESSAY: GLOBAL HEALTH RELEVANCE

Substance use disorders are a growing burden in Nigeria and across sub-Saharan Africa. The World Health Organization estimates that fewer than one in five people with substance use disorders in Africa receive any treatment. One reason is the lack of computational tools that can guide treatment selection at the individual level. My CCT model addresses this directly by providing a quantitative framework for predicting which pharmacological combinations will prevent reward-memory consolidation.

The model is designed to be drug-agnostic and species-agnostic. It can be parameterized for any compound with known receptor binding profiles and any species with known dopamine, NMDA, and affective processing dynamics. This means it can be applied to the specific drugs and populations relevant to Nigeria, where treatment resources are limited and trial-and-error prescribing is the norm.

My background as a Nigerian pharmacist and independent researcher gives me direct insight into the clinical realities of addiction treatment in low-resource settings. I have seen patients cycle through treatments without any quantitative guidance. The CCT model, once validated against behavioral data, could be deployed as a clinical decision support tool. The fellowship's U.S. visit component is essential for accessing the experimental data needed for that validation.

SHORT ESSAY: MENTORSHIP AND COLLABORATION GOALS

I seek mentorship in two specific areas. First, fitting the CCT model and neurocascade engine to real behavioral data from animal models of addiction. My models are calibrated to literature values, but they have not been fitted to raw behavioral time series. I need a mentor with expertise in computational neuroscience and access to rodent operant conditioning datasets. Second, translating the validated model into a clinically usable tool, which requires expertise in digital health implementation and regulatory pathways.

I have already established connections with Kent Berridge at the University of Michigan, Samuel Gershman at Harvard, Nathaniel Daw at Princeton, and Marcelo Mattar at NYU. Any of these groups would be a suitable host. I am open to other institutions with strong computational neuroscience and addiction research programmes.

The fellowship would fund a three-to-six-month visit. During that time, I would complete the model fitting, write a manuscript for submission to a peer-reviewed journal, and develop a prototype clinical decision support tool. I would also give a seminar at the host institution and participate in lab meetings and journal clubs. After returning to Nigeria, I would continue the collaboration remotely and apply for follow-on funding to deploy the tool in a clinical pilot study.

CHECKLIST

- [ ] Motivation letter, 500 words maximum
- [ ] Research statement, 600 words maximum
- [ ] Short essay on global health relevance, 350 words maximum
- [ ] Short essay on mentorship and collaboration goals, 350 words maximum
- [ ] Curriculum vitae with ORCID, GitHub, and personal website
- [ ] Academic transcripts from University of Ibadan and Hasso Plattner Institute
- [ ] Two letters of recommendation from endorsers listed in profile
- [ ] Pre-registration documents for CCT model and neurocascade studies
- [ ] Copies of three sole-authored preprints on OSF or Zenodo
- [ ] Copy of co-authored paper under review at Alcohol
- [ ] Proof of Nigerian citizenship
- [ ] Proof of enrollment in M.Sc. Digital Health at Hasso Plattner Institute
- [ ] Proposed U.S. host institution letter of support
- [ ] Budget outline for three-to-six-month U.S. visit

EDITOR NOTES

- Eligibility risk: The fellowship may require a PhD or equivalent research experience. Eniola has a B.Pharm and is enrolled in an M.Sc. but does not hold a PhD. Confirm that the programme accepts pre-PhD candidates or that the M.Sc. enrollment satisfies the requirement.
- Fact to verify: The fellowship website URL provided may have changed. Confirm the current URL and any updates to eligibility criteria or deadlines.
- Gap: The profile does not specify which U.S. host institution Eniola would propose. The applicant must identify a specific lab or mentor and obtain a letter of support before submission.
- Gap: The profile does not include a budget estimate for the U.S. visit. The applicant should calculate travel, accommodation, and living expenses for three to six months.
- Fact to verify: The co-authored paper under review at Alcohol may have been accepted or rejected by the time of submission. Confirm the current status and update the application accordingly.

MOTIVATION LETTER

Substance use disorders are rising across sub-Saharan Africa, yet the computational tools to model their neurobiological drivers remain almost entirely developed in and for high-income settings. My research directly addresses this gap. I am Eniola Ayodele Olutogun, an independent computational researcher and licensed pharmacist from Nigeria, and I am applying to the Global Health Emerging Scholars Fellowship to advance my Conjunctive Consolidation Threshold model of reward-memory encoding in addiction.

The CCT model is a tripartite pharmacological framework I developed and tested entirely as an independent researcher. It couples dopaminergic reward-prediction-error signals, NMDAR-dependent long-term potentiation, and affective contrast into a single ordinary differential equation system. I calibrated the model using Bayesian MCMC with literature-elicited priors drawn from a systematic screen of 1,847 records. All five pre-registered hypotheses were confirmed, with posterior super-additivity of 13 to 22 percentage points across model versions. Three sole-authored preprints are deposited on OSF and Zenodo, and a co-authored paper is under review at Alcohol.

This fellowship would allow me to spend three to six months at a U.S. host institution, working with a mentor to fit the CCT model to real behavioral data from human or animal studies. My neurocascade simulation engine, which couples pharmacokinetics to receptor-binding to Wilson-Cowan circuit dynamics, is ready to serve as the translational bridge. The engine passes 62 of 62 tests and is calibrated for mu-opioid, D2 dopamine, and GABA-A receptor systems. A U.S. visit would provide access to experimental collaborators and datasets that are not available in Nigeria.

My academic record includes a B.Pharm from the University of Ibadan with a German-equivalent grade of 1.9, and I am enrolled in the M.Sc. Digital Health programme at the Hasso Plattner Institute in Potsdam, Germany, starting winter 2026. My independent research output includes pre-registered studies on hERG cardiotoxicity topology, drug-resistance mutation prediction using protein-language models, and Koopman-operator methods for drug transport in tumor tissue. I have endorsements from Kent Berridge at Michigan, Samuel Gershman at Harvard, Nathaniel Daw at Princeton, and Marcelo Mattar at NYU.

The Global Health Emerging Scholars Fellowship is the right programme for this work because it explicitly supports early-career researchers from LMICs, values rigorous computational approaches, and requires a U.S. visit that would directly enable the collaboration my models need. I am ready to commit to that visit and to return to Nigeria with a clinically actionable computational framework for addiction neuroscience.

RESEARCH STATEMENT

My research programme addresses a specific global health problem: the lack of computational models that can predict, at the individual level, whether a given pharmacological intervention will prevent the consolidation of reward memories in substance use disorders. The standard approach in addiction neuroscience relies on qualitative circuit diagrams and single-pathway hypotheses. I build quantitative, testable models that integrate multiple neurotransmitter systems and produce falsifiable predictions.

The CCT model is the core of this programme. It posits that reward-memory consolidation requires the conjunctive activation of three axes: a dopaminergic reward-prediction-error signal above a threshold, NMDAR-dependent long-term potentiation above a separate threshold, and a positive affective contrast signal. The model is implemented as a system of coupled ODEs solved with RK45 and calibrated with Bayesian MCMC using PyMC's DEMetropolisZ sampler. The 14 free parameters were assigned priors from a systematic literature screen of 1,847 records. All five pre-registered hypotheses were confirmed. The model predicts that partial blockade of any single axis is insufficient; only simultaneous reduction of all three below their thresholds prevents consolidation. This has direct implications for combination pharmacotherapy in addiction.

The neurocascade simulation engine extends this work by modeling the full chain from drug pharmacokinetics to receptor binding to circuit dynamics to behavioral readout. It is implemented in Python with ODE layers for each stage and calibrated with Bayesian methods for three receptor systems. All 62 unit tests pass. The circuit-layer parameters are currently illustrative, pending fits to real behavioral data. That is the next step, and it requires access to experimental collaborators and datasets.

My other research lines demonstrate the same methodological approach. The hERG cardiotoxicity study was a pre-registered, powered replication that tested whether bipartite persistent homology could predict cardiotoxicity from protein-ligand interface geometry. It found that topological features do not beat a plain descriptor baseline, settling a comparison the literature had never actually run. The interface-topology-for-resistance study applied the same methods to drug-resistance prediction and found they carry almost no signal, ruling out interface geometry as the driver. The TOPOLOGIX project then built a sequence-based predictor using ESM-2 protein-language-model delta-embeddings and Morgan fingerprints, achieving AUROC 0.804 on the Platinum benchmark and covering 100% of mutations versus approximately 18% for structure-limited tools.

The fellowship would fund a three-to-six-month visit to a U.S. host institution where I would fit the CCT model and neurocascade engine to behavioral data from animal models of addiction. The expected outcome is a validated, open-source computational framework that predicts the efficacy of combination pharmacotherapies for substance use disorders. This framework would be directly applicable to the Nigerian context, where opioid and methamphetamine use are rising and where computational pharmacology expertise is scarce.

SHORT ESSAY: GLOBAL HEALTH RELEVANCE

Substance use disorders are a growing burden in Nigeria and across sub-Saharan Africa. The World Health Organization estimates that fewer than one in five people with substance use disorders in Africa receive any treatment. One reason is the lack of computational tools that can guide treatment selection at the individual level. My CCT model addresses this directly by providing a quantitative framework for predicting which pharmacological combinations will prevent reward-memory consolidation.

The model is designed to be drug-agnostic and species-agnostic. It can be parameterized for any compound with known receptor binding profiles and any species with known dopamine, NMDA, and affective processing dynamics. This means it can be applied to the specific drugs and populations relevant to Nigeria, where treatment resources are limited and trial-and-error prescribing is the norm.

My background as a Nigerian pharmacist and independent researcher gives me direct insight into the clinical realities of addiction treatment in low-resource settings. I have seen patients cycle through treatments without any quantitative guidance. The CCT model, once validated against behavioral data, could be deployed as a clinical decision support tool. The fellowship's U.S. visit component is essential for accessing the experimental data needed for that validation.

SHORT ESSAY: MENTORSHIP AND COLLABORATION GOALS

I seek mentorship in two specific areas. First, fitting the CCT model and neurocascade engine to real behavioral data from animal models of addiction. My models are calibrated to literature values, but they have not been fitted to raw behavioral time series. I need a mentor with expertise in computational neuroscience and access to rodent operant conditioning datasets. Second, translating the validated model into a clinically usable tool, which requires expertise in digital health implementation and regulatory pathways.

I have already established connections with Kent Berridge at the University of Michigan, Samuel Gershman at Harvard, Nathaniel Daw at Princeton, and Marcelo Mattar at NYU. Any of these groups would be a suitable host. I am open to other institutions with strong computational neuroscience and addiction research programmes.

The fellowship would fund a three-to-six-month visit. During that time, I would complete the model fitting, write a manuscript for submission to a peer-reviewed journal, and develop a prototype clinical decision support tool. I would also give a seminar at the host institution and participate in lab meetings and journal clubs. After returning to Nigeria, I would continue the collaboration remotely and apply for follow-on funding to deploy the tool in a clinical pilot study.

CHECKLIST

- [ ] Motivation letter, 500 words maximum
- [ ] Research statement, 600 words maximum
- [ ] Short essay on global health relevance, 350 words maximum
- [ ] Short essay on mentorship and collaboration goals, 350 words maximum
- [ ] Curriculum vitae with ORCID, GitHub, and personal website
- [ ] Academic transcripts from University of Ibadan and Hasso Plattner Institute
- [ ] Two letters of recommendation from endorsers listed in profile
- [ ] Pre-registration documents for CCT model and neurocascade studies
- [ ] Copies of three sole-authored preprints on OSF or Zenodo
- [ ] Copy of co-authored paper under review at Alcohol
- [ ] Proof of Nigerian citizenship
- [ ] Proof of enrollment in M.Sc. Digital Health at Hasso Plattner Institute
- [ ] Proposed U.S. host institution letter of support
- [ ] Budget outline for three-to-six-month U.S. visit

EDITOR NOTES

- Eligibility risk: The fellowship may require a PhD or equivalent research experience. Eniola has a B.Pharm and is enrolled in an M.Sc. but does not hold a PhD. Confirm that the programme accepts pre-PhD candidates or that the M.Sc. enrollment satisfies the requirement.
- Fact to verify: The fellowship website URL provided may have changed. Confirm the current URL and any updates to eligibility criteria or deadlines.
- Gap: The profile does not specify which U.S. host institution Eniola would propose. The applicant must identify a specific lab or mentor and obtain a letter of support before submission.
- Gap: The profile does not include a budget estimate for the U.S. visit. The applicant should calculate travel, accommodation, and living expenses for three to six months.
- Fact to verify: The co-authored paper under review at Alcohol may have been accepted or rejected by the time of submission. Confirm the current status and update the application accordingly.

AI Draft — Global Health Emerging Scholars Fellowship