Draft — CIFAR Azrieli Global Scholars Program

MOTIVATION LETTER

The CIFAR Azrieli Global Scholars Program targets early-career researchers building independent programs. I am an independent researcher in Lagos, Nigeria, and I have built one. Between January 2025 and March 2026, I designed, mathematically specified, and computationally validated the Conjunctive Consolidation Threshold (CCT) model, a tripartite pharmacological framework for preventing reward-memory encoding in addiction. The model reduced encoding probability from 0.855 to 0.122, an 85.8 percent reduction, with super-additivity of 12.8 percentage points. All five pre-registered hypotheses H1 through H5 were confirmed. I published three sole-authored preprints on OSF and Zenodo, and a review article is under review at Neuroscience and Biobehavioral Reviews. I built three open-source platforms: IMPRINT for addiction-liability screening, TOPOLOGIX for topological data analysis of drug-protein interactions, and GATE for BCI neural-stimulation safety evaluation. I hold a provisional patent on the CCT core architecture, filed Q3 2026. My work has received endorsements from Kent Berridge at Michigan, Samuel Gershman at Harvard, Nathaniel Daw at Princeton, and Marcelo Mattar at NYU.

The CCT model aligns directly with the CIFAR program Learning in Machines and Brains. The model formalizes how conjunctive representations of drug reward and environmental context are consolidated into long-term memory, and it proposes a pharmacological intervention at the threshold of that consolidation. This is a computational neuroscience question with a pharmacological answer. I validated the model using ODE/RK45 numerical integration and Bayesian MCMC population dynamics. I am now extending the model to account for individual variability in addiction liability using Bayesian hierarchical models, and I am designing a clinical trial architecture for a three-arm, double-blind, placebo-controlled study.

I currently serve as National Product Manager at Synthcare in Lagos, a role that funds my independent research. I manage product strategy for pharmaceutical distribution across Nigeria. This is a full-time position, and I treat my research program as equivalent to an academic appointment. I supervise no trainees formally, but I mentor two junior researchers at ZYCO, my independent research collective, and I maintain active collaborations with the four senior scientists named above. I hold a B.Pharm from the University of Ibadan with a CGPA of 5.1 out of 7.0, a German equivalent of 1.9, and I am a PCN-licensed pharmacist. I am applying to MSc programs in computational neuroscience at MUG and Graz University of Technology for an October 2026 start.

The CIFAR network would accelerate the CCT model from computational validation to clinical trial design. I need access to experimental collaborators who can test the model in animal models of addiction, and I need exposure to the machine learning and theoretical neuroscience communities within Learning in Machines and Brains. The travel support would allow me to attend CIFAR meetings and present my work at international conferences. I hold a valid Nigerian passport and can travel internationally two to three times per year. I am available for a virtual interview on March 24 through 26, 2026.

RESEARCH STATEMENT

The Conjunctive Consolidation Threshold (CCT) model addresses a fundamental gap in addiction neuroscience: how drug reward and environmental context are bound together into a persistent memory trace, and how that binding can be prevented pharmacologically. Existing models of addiction focus on dopamine-mediated reward prediction error, habit formation in the dorsolateral striatum, or incentive salience attribution to drug cues. None of these models specify the computational threshold at which a conjunctive representation of reward and context transitions from short-term to long-term storage. The CCT model fills that gap.

The model has three components. First, a conjunctive encoding stage in which hippocampal place cells and ventral tegmental area dopamine neurons co-activate to bind drug reward to spatial and temporal context. Second, a consolidation threshold determined by the strength and synchrony of that co-activation, modeled as a sigmoidal function of dopamine concentration and NMDA receptor conductance. Third, a pharmacological intervention that raises the threshold by partial NMDA receptor antagonism combined with dopamine D1 receptor modulation, preventing consolidation without blocking acute reward. The mathematical specification is published on OSF (DOI 10.17605/OSF.IO/EMY4U). The Bayesian population dynamics and clinical trial architecture are published on Zenodo (DOI 10.5281/zenodo.20492472).

I validated the model computationally using ODE/RK45 numerical integration in Python with scipy and numpy, and Bayesian MCMC parameter estimation using PyMC. The encoding probability dropped from 0.855 to 0.122 under the combined intervention, an 85.8 percent reduction. The combination showed super-additivity of 12.8 percentage points, meaning the combined effect exceeded the sum of the individual effects. I confirmed all five pre-registered hypotheses: H1 (conjunctive encoding requires synchronous DA and NMDA activation), H2 (threshold is sigmoidal), H3 (partial NMDA antagonism raises threshold), H4 (D1 modulation raises threshold), and H5 (combination is super-additive). The model is now under peer review as a review article at Neuroscience and Biobehavioral Reviews.

I built three platforms to support the CCT model and related work. IMPRINT screens compounds for addiction liability by simulating their effect on the CCT threshold. TOPOLOGIX uses topological data analysis with persistent homology and bipartite simplicial complexes to predict drug-protein interactions, with a validated MVP for hERG cardiotoxicity. GATE evaluates safety of BCI neural-stimulation protocols using biophysical neuron models in NEURON and Brian2. All three are open-source under Apache 2.0.

The next phase of the CCT project has three aims. First, extend the model to a Bayesian hierarchical framework that accounts for individual variability in dopamine transporter density, NMDA receptor subunit composition, and baseline memory consolidation rates. Second, design a three-arm, double-blind, placebo-controlled clinical trial with 120 participants, using the Bayesian population dynamics to determine sample size and power. Third, collaborate with experimental neuroscientists to test the model in rodent models of context-induced reinstatement. The CIFAR Learning in Machines and Brains program is the ideal environment for this work because it brings together computational modelers, experimentalists, and clinicians who study learning and memory at multiple levels of analysis.

I hold a provisional patent on the CCT core architecture, filed Q3 2026. I have endorsements from Kent Berridge, Samuel Gershman, Nathaniel Daw, and Marcelo Mattar. I am an independent researcher based in Lagos, Nigeria, and I am committed to building computational neuroscience capacity in Africa. The CCT model is the foundation of my independent research program, and the CIFAR Azrieli Global Scholars Program would provide the network, funding, and credibility to move it from computational validation to clinical application.

SHORT ESSAY: LEADERSHIP AND INDEPENDENCE

I lead an independent research program from Lagos, Nigeria, without a PhD or a formal academic appointment. I designed the CCT model, wrote all three preprints, built three open-source platforms, and filed a provisional patent, all between January 2025 and March 2026. I manage my research timeline, budget, and collaborations independently. I fund my research through my salary as National Product Manager at Synthcare, a role that requires me to manage product strategy, supply chain logistics, and cross-functional teams across Nigeria. That role has given me project management, budgeting, and stakeholder communication skills that transfer directly to leading a research program.

I have built a network of senior collaborators who endorse my work. Kent Berridge, Samuel Gershman, Nathaniel Daw, and Marcelo Mattar have all reviewed my preprints and provided feedback. Gershman endorsed my arXiv submission. These relationships are peer collaborations between an independent researcher and senior scientists. I maintain them through regular email correspondence, shared code repositories, and joint manuscript preparation. I am co-author on a paper under review at Alcohol (Elsevier).

I mentor two junior researchers at ZYCO, my independent research collective. One is a final-year pharmacy student learning Python and Bayesian statistics. The other is a recent graduate building a QSAR model for NMDA receptor ligands. I teach them through weekly code reviews, shared reading groups, and collaborative coding sessions. This is not formal supervision, but it is training of the next generation of computational neuroscientists in Nigeria.

SHORT ESSAY: FIT WITH CIFAR AND LEARNING IN MACHINES AND BRAINS

The CCT model is a computational framework for learning and memory that bridges machine learning and neuroscience. The model formalizes conjunctive representation learning, a concept from hippocampal-dependent episodic memory, as a thresholded consolidation process. This is directly relevant to the Learning in Machines and Brains program, which studies how learning algorithms in biological and artificial systems can inform each other. My model proposes a specific algorithm for when and how conjunctive representations are stored, and it suggests a pharmacological intervention that modulates that algorithm. This is a concrete example of how computational principles can guide therapeutic design.

I have the technical skills to engage with the machine learning community. I use Python with scipy, numpy, PyMC, and ODE solvers. I have built Bayesian population dynamics models and validated them with MCMC. I have experience with topological data analysis using Ripser and Gudhi. I have built web applications with JavaScript and Node.js, and I manage databases with Supabase and Postgres. I can contribute to discussions about representation learning, Bayesian inference, and dynamical systems in the context of learning and memory.

The CIFAR network would give me access to experimental collaborators who can test the CCT model in animal models. I need rodent behavior data to validate the model's predictions about context-induced reinstatement. I need human neuroimaging data to test the model's predictions about hippocampal and striatal activation during drug-context encoding. The Learning in Machines and Brains program includes experimentalists who collect these data. I can provide the computational framework; they can provide the empirical tests. This is the collaboration I need to move the CCT model from theory to application.

CHECKLIST

- [ ] Confirm eligibility: PhD or equivalent degree held by application deadline. Eniola holds a B.Pharm, not a PhD. Verify if CIFAR accepts B.Pharm plus independent research record as equivalent.
- [ ] Confirm eligibility: Full-time academic position with independent research program and trainee supervision, started between July 1, 2021 and December 31, 2025, and secured through March 31, 2028. Eniola is a National Product Manager at Synthcare, not an academic. Verify if CIFAR accepts industry positions with independent research programs.
- [ ] Confirm eligibility: Post-doctoral fellows are not eligible. Eniola is not a post-doc, so this condition is met.
- [ ] Confirm availability for virtual interview on March 24-26, 2026.
- [ ] Confirm valid passport and ability to travel internationally 2-3 times per year.
- [ ] Prepare CV or resume in CIFAR format.
- [ ] Prepare two letters of reference. Candidates: Kent Berridge, Samuel Gershman, Nathaniel Daw, or Marcelo Mattar.
- [ ] Prepare research proposal summary (if required by CIFAR application portal).
- [ ] Submit application through CIFAR online portal before deadline.
- [ ] Verify deadline on CIFAR website.

EDITOR NOTES

- Eligibility risk: The CIFAR Azrieli Global Scholars Program requires a PhD or equivalent and a full-time academic appointment. Eniola has neither. The application may be rejected at the eligibility screening stage. Consider contacting CIFAR program officer to ask whether independent researchers with industry positions and strong publication records are considered equivalent. If not, this application may not be viable.
- Verification needed: Confirm that the CIFAR program is recruiting for Learning in Machines and Brains in the 2026-2028 cycle. The program rotates which themes recruit each year. Check the CIFAR website.
- Gap: The profile does not include a date of birth or exact age. The profile says age 29, but the application may require a birth date. Insert Eniola's actual birth date.
- Gap: The profile does not include a list of publications with full citations. The application may require a publication list. Prepare a formatted list including the three preprints, the review article under review, and the co-authored paper under review at Alcohol.
- Gap: The profile does not include a detailed budget or justification for the $10,000 per year. Prepare a budget that covers travel to CIFAR meetings, conference registration, and computational resources (cloud computing for Bayesian MCMC, GPU time for TDA).

MOTIVATION LETTER

The CIFAR Azrieli Global Scholars Program targets early-career researchers building independent programs. I am an independent researcher in Lagos, Nigeria, and I have built one. Between January 2025 and March 2026, I designed, mathematically specified, and computationally validated the Conjunctive Consolidation Threshold (CCT) model, a tripartite pharmacological framework for preventing reward-memory encoding in addiction. The model reduced encoding probability from 0.855 to 0.122, an 85.8 percent reduction, with super-additivity of 12.8 percentage points. All five pre-registered hypotheses H1 through H5 were confirmed. I published three sole-authored preprints on OSF and Zenodo, and a review article is under review at Neuroscience and Biobehavioral Reviews. I built three open-source platforms: IMPRINT for addiction-liability screening, TOPOLOGIX for topological data analysis of drug-protein interactions, and GATE for BCI neural-stimulation safety evaluation. I hold a provisional patent on the CCT core architecture, filed Q3 2026. My work has received endorsements from Kent Berridge at Michigan, Samuel Gershman at Harvard, Nathaniel Daw at Princeton, and Marcelo Mattar at NYU.

The CCT model aligns directly with the CIFAR program Learning in Machines and Brains. The model formalizes how conjunctive representations of drug reward and environmental context are consolidated into long-term memory, and it proposes a pharmacological intervention at the threshold of that consolidation. This is a computational neuroscience question with a pharmacological answer. I validated the model using ODE/RK45 numerical integration and Bayesian MCMC population dynamics. I am now extending the model to account for individual variability in addiction liability using Bayesian hierarchical models, and I am designing a clinical trial architecture for a three-arm, double-blind, placebo-controlled study.

I currently serve as National Product Manager at Synthcare in Lagos, a role that funds my independent research. I manage product strategy for pharmaceutical distribution across Nigeria. This is a full-time position, and I treat my research program as equivalent to an academic appointment. I supervise no trainees formally, but I mentor two junior researchers at ZYCO, my independent research collective, and I maintain active collaborations with the four senior scientists named above. I hold a B.Pharm from the University of Ibadan with a CGPA of 5.1 out of 7.0, a German equivalent of 1.9, and I am a PCN-licensed pharmacist. I am applying to MSc programs in computational neuroscience at MUG and Graz University of Technology for an October 2026 start.

The CIFAR network would accelerate the CCT model from computational validation to clinical trial design. I need access to experimental collaborators who can test the model in animal models of addiction, and I need exposure to the machine learning and theoretical neuroscience communities within Learning in Machines and Brains. The travel support would allow me to attend CIFAR meetings and present my work at international conferences. I hold a valid Nigerian passport and can travel internationally two to three times per year. I am available for a virtual interview on March 24 through 26, 2026.

RESEARCH STATEMENT

The Conjunctive Consolidation Threshold (CCT) model addresses a fundamental gap in addiction neuroscience: how drug reward and environmental context are bound together into a persistent memory trace, and how that binding can be prevented pharmacologically. Existing models of addiction focus on dopamine-mediated reward prediction error, habit formation in the dorsolateral striatum, or incentive salience attribution to drug cues. None of these models specify the computational threshold at which a conjunctive representation of reward and context transitions from short-term to long-term storage. The CCT model fills that gap.

The model has three components. First, a conjunctive encoding stage in which hippocampal place cells and ventral tegmental area dopamine neurons co-activate to bind drug reward to spatial and temporal context. Second, a consolidation threshold determined by the strength and synchrony of that co-activation, modeled as a sigmoidal function of dopamine concentration and NMDA receptor conductance. Third, a pharmacological intervention that raises the threshold by partial NMDA receptor antagonism combined with dopamine D1 receptor modulation, preventing consolidation without blocking acute reward. The mathematical specification is published on OSF (DOI 10.17605/OSF.IO/EMY4U). The Bayesian population dynamics and clinical trial architecture are published on Zenodo (DOI 10.5281/zenodo.20492472).

I validated the model computationally using ODE/RK45 numerical integration in Python with scipy and numpy, and Bayesian MCMC parameter estimation using PyMC. The encoding probability dropped from 0.855 to 0.122 under the combined intervention, an 85.8 percent reduction. The combination showed super-additivity of 12.8 percentage points, meaning the combined effect exceeded the sum of the individual effects. I confirmed all five pre-registered hypotheses: H1 (conjunctive encoding requires synchronous DA and NMDA activation), H2 (threshold is sigmoidal), H3 (partial NMDA antagonism raises threshold), H4 (D1 modulation raises threshold), and H5 (combination is super-additive). The model is now under peer review as a review article at Neuroscience and Biobehavioral Reviews.

I built three platforms to support the CCT model and related work. IMPRINT screens compounds for addiction liability by simulating their effect on the CCT threshold. TOPOLOGIX uses topological data analysis with persistent homology and bipartite simplicial complexes to predict drug-protein interactions, with a validated MVP for hERG cardiotoxicity. GATE evaluates safety of BCI neural-stimulation protocols using biophysical neuron models in NEURON and Brian2. All three are open-source under Apache 2.0.

The next phase of the CCT project has three aims. First, extend the model to a Bayesian hierarchical framework that accounts for individual variability in dopamine transporter density, NMDA receptor subunit composition, and baseline memory consolidation rates. Second, design a three-arm, double-blind, placebo-controlled clinical trial with 120 participants, using the Bayesian population dynamics to determine sample size and power. Third, collaborate with experimental neuroscientists to test the model in rodent models of context-induced reinstatement. The CIFAR Learning in Machines and Brains program is the ideal environment for this work because it brings together computational modelers, experimentalists, and clinicians who study learning and memory at multiple levels of analysis.

I hold a provisional patent on the CCT core architecture, filed Q3 2026. I have endorsements from Kent Berridge, Samuel Gershman, Nathaniel Daw, and Marcelo Mattar. I am an independent researcher based in Lagos, Nigeria, and I am committed to building computational neuroscience capacity in Africa. The CCT model is the foundation of my independent research program, and the CIFAR Azrieli Global Scholars Program would provide the network, funding, and credibility to move it from computational validation to clinical application.

SHORT ESSAY: LEADERSHIP AND INDEPENDENCE

I lead an independent research program from Lagos, Nigeria, without a PhD or a formal academic appointment. I designed the CCT model, wrote all three preprints, built three open-source platforms, and filed a provisional patent, all between January 2025 and March 2026. I manage my research timeline, budget, and collaborations independently. I fund my research through my salary as National Product Manager at Synthcare, a role that requires me to manage product strategy, supply chain logistics, and cross-functional teams across Nigeria. That role has given me project management, budgeting, and stakeholder communication skills that transfer directly to leading a research program.

I have built a network of senior collaborators who endorse my work. Kent Berridge, Samuel Gershman, Nathaniel Daw, and Marcelo Mattar have all reviewed my preprints and provided feedback. Gershman endorsed my arXiv submission. These relationships are peer collaborations between an independent researcher and senior scientists. I maintain them through regular email correspondence, shared code repositories, and joint manuscript preparation. I am co-author on a paper under review at Alcohol (Elsevier).

I mentor two junior researchers at ZYCO, my independent research collective. One is a final-year pharmacy student learning Python and Bayesian statistics. The other is a recent graduate building a QSAR model for NMDA receptor ligands. I teach them through weekly code reviews, shared reading groups, and collaborative coding sessions. This is not formal supervision, but it is training of the next generation of computational neuroscientists in Nigeria.

SHORT ESSAY: FIT WITH CIFAR AND LEARNING IN MACHINES AND BRAINS

The CCT model is a computational framework for learning and memory that bridges machine learning and neuroscience. The model formalizes conjunctive representation learning, a concept from hippocampal-dependent episodic memory, as a thresholded consolidation process. This is directly relevant to the Learning in Machines and Brains program, which studies how learning algorithms in biological and artificial systems can inform each other. My model proposes a specific algorithm for when and how conjunctive representations are stored, and it suggests a pharmacological intervention that modulates that algorithm. This is a concrete example of how computational principles can guide therapeutic design.

I have the technical skills to engage with the machine learning community. I use Python with scipy, numpy, PyMC, and ODE solvers. I have built Bayesian population dynamics models and validated them with MCMC. I have experience with topological data analysis using Ripser and Gudhi. I have built web applications with JavaScript and Node.js, and I manage databases with Supabase and Postgres. I can contribute to discussions about representation learning, Bayesian inference, and dynamical systems in the context of learning and memory.

The CIFAR network would give me access to experimental collaborators who can test the CCT model in animal models. I need rodent behavior data to validate the model's predictions about context-induced reinstatement. I need human neuroimaging data to test the model's predictions about hippocampal and striatal activation during drug-context encoding. The Learning in Machines and Brains program includes experimentalists who collect these data. I can provide the computational framework; they can provide the empirical tests. This is the collaboration I need to move the CCT model from theory to application.

CHECKLIST

- [ ] Confirm eligibility: PhD or equivalent degree held by application deadline. Eniola holds a B.Pharm, not a PhD. Verify if CIFAR accepts B.Pharm plus independent research record as equivalent.
- [ ] Confirm eligibility: Full-time academic position with independent research program and trainee supervision, started between July 1, 2021 and December 31, 2025, and secured through March 31, 2028. Eniola is a National Product Manager at Synthcare, not an academic. Verify if CIFAR accepts industry positions with independent research programs.
- [ ] Confirm eligibility: Post-doctoral fellows are not eligible. Eniola is not a post-doc, so this condition is met.
- [ ] Confirm availability for virtual interview on March 24-26, 2026.
- [ ] Confirm valid passport and ability to travel internationally 2-3 times per year.
- [ ] Prepare CV or resume in CIFAR format.
- [ ] Prepare two letters of reference. Candidates: Kent Berridge, Samuel Gershman, Nathaniel Daw, or Marcelo Mattar.
- [ ] Prepare research proposal summary (if required by CIFAR application portal).
- [ ] Submit application through CIFAR online portal before deadline.
- [ ] Verify deadline on CIFAR website.

EDITOR NOTES

- Eligibility risk: The CIFAR Azrieli Global Scholars Program requires a PhD or equivalent and a full-time academic appointment. Eniola has neither. The application may be rejected at the eligibility screening stage. Consider contacting CIFAR program officer to ask whether independent researchers with industry positions and strong publication records are considered equivalent. If not, this application may not be viable.
- Verification needed: Confirm that the CIFAR program is recruiting for Learning in Machines and Brains in the 2026-2028 cycle. The program rotates which themes recruit each year. Check the CIFAR website.
- Gap: The profile does not include a date of birth or exact age. The profile says age 29, but the application may require a birth date. Insert Eniola's actual birth date.
- Gap: The profile does not include a list of publications with full citations. The application may require a publication list. Prepare a formatted list including the three preprints, the review article under review, and the co-authored paper under review at Alcohol.
- Gap: The profile does not include a detailed budget or justification for the $10,000 per year. Prepare a budget that covers travel to CIFAR meetings, conference registration, and computational resources (cloud computing for Bayesian MCMC, GPU time for TDA).

AI Draft — CIFAR Azrieli Global Scholars Program