MOTIVATION LETTER The Conjunctive Consolidation Threshold model, developed independently over the past eighteen months in Lagos, proposes that reward-memory encoding in addiction follows a tripartite pharmacological threshold that can be mathematically specified and clinically interrupted. Three sole-authored preprints on OSF and Zenodo formalize this framework: the foundational CCT paper, a complete mathematical specification using ODE/RK45 systems, and a Bayesian population dynamics model with a clinical trial architecture. Validation against pre-registered hypotheses H1 through H5 produced an 85.8 percent reduction in encoding probability, from 0.855 to 0.122, with super-additivity of 12.8 percentage points. This work has received endorsements from Kent Berridge at Michigan, Samuel Gershman at Harvard, Nathaniel Daw at Princeton, and Marcelo Mattar at NYU. A provisional patent on the CCT core architecture is filed for Q3 2026. This PhD Computational Neuroscience Fellowship is the precise mechanism to formalize the CCT model within a top-tier computational neuroscience laboratory, extending the mathematical framework into spiking neural network simulations and clinical trial design. The programme's emphasis on computational methods and neuroscience knowledge directly matches my demonstrated proficiency in ODE/RK45, Bayesian MCMC with PyMC, topological data analysis with Ripser and Gudhi, and neural simulation with NEURON and Brian2. My B.Pharm from the University of Ibadan with a German equivalent grade of 1.9 provides the pharmacological foundation; the fellowship would supply the computational neuroscience depth that my independent work has already begun to build. Nigeria has no dedicated computational neuroscience PhD programme. Every line of code for the CCT model was written on personal hardware in Lagos, with no institutional supercomputing access. The fellowship would enable formal training in neural data analysis, reinforcement learning theory, and population dynamics that I have taught myself through arXiv preprints and open-source repositories. My goal is to return to West Africa with the credentials and network to establish a computational addiction neuroscience group, addressing a substance-use burden that existing pharmacological models have failed to reduce. The selection criteria for this fellowship emphasize academic excellence, research potential, technical skills, and fit with the programme. My CGPA of 5.1 out of 7.0, three sole-authored preprints, one review article under review at Neuroscience and Biobehavioral Reviews, and a co-authored paper under review at Alcohol demonstrate the first two criteria. The technical stack listed on my GitHub and ORCID profiles satisfies the third. The fourth criterion, letters of recommendation, can be supplied by the four named endorsers who have reviewed the CCT mathematical specification. The fifth criterion, clarity of career goals, is stated above: formalize the CCT model, publish in high-impact computational neuroscience journals, and build African research capacity. RESEARCH STATEMENT The Conjunctive Consolidation Threshold model addresses a specific gap in addiction neuroscience: no existing pharmacological framework predicts the precise moment at which a reward experience becomes consolidated into a long-term memory trace that drives compulsive drug-seeking. Current models treat reward processing and memory consolidation as separate systems. The CCT model proposes that three concurrent conditions must cross a conjunctive threshold for encoding to occur: dopaminergic salience, glutamatergic plasticity permissiveness, and cholinergic timing coherence. When any one condition falls below threshold, encoding probability drops to near-zero. The mathematical specification uses a system of coupled ordinary differential equations solved via Runge-Kutta 45 integration, with parameters estimated through Bayesian Markov chain Monte Carlo sampling. The population dynamics extension models inter-individual variability in threshold values across a simulated cohort of 10,000 agents, producing the 85.8 percent reduction in encoding probability that was confirmed against all five pre-registered hypotheses. The super-additivity result of 12.8 percentage points indicates that the three conditions interact non-linearly, which standard pharmacological models cannot capture. During this fellowship, I propose to extend the CCT model in three directions. First, implement the threshold dynamics in a spiking neural network using Brian2, simulating dopaminergic, glutamatergic, and cholinergic inputs to a hippocampal-striatal circuit. Second, validate the model against existing rodent self-administration datasets available through the Collaborative Research in Computational Neuroscience data sharing initiative. Third, design a Bayesian adaptive clinical trial protocol that uses the CCT framework to stratify patients by predicted encoding probability, enabling personalized intervention timing. The technical skills required for this work are already demonstrated. I have built three open-source platforms: IMPRINT for addiction-liability screening, TOPOLOGIX for topological data analysis of drug-protein interactions using persistent homology and bipartite simplicial complexes, and GATE for BCI neural-stimulation safety evaluation under Apache 2.0. The hERG cardiotoxicity MVP within TOPOLOGIX shows my ability to translate computational methods to pharmacological safety questions. My bioinformatics work with GHRU-GSAR on antimicrobial resistance genomics and surveillance pipelines demonstrates experience with high-performance computing workflows using Nextflow and SLURM. The fellowship would provide formal training in neural data analysis methods, reinforcement learning theory as applied to addiction, and advanced Bayesian statistics, areas where my self-taught knowledge requires structured deepening. The expected output is a series of three first-author papers: the spiking neural network validation, the clinical trial protocol, and a review of computational approaches to addiction pharmacology. The provisional patent provides a pathway toward translational application. SHORT ESSAY: MOTIVATION FOR COMPUTATIONAL NEUROSCIENCE Pharmacology gave me the molecular mechanism. Computational neuroscience gives me the dynamical system. During my B.Pharm at the University of Ibadan, I learned that drugs bind to receptors, but I could not explain why two patients with identical receptor profiles respond differently to the same addiction treatment. The answer, I now understand, lies in the population dynamics of neural circuits, not in the equilibrium binding constant. The CCT model emerged from this realization: addiction is a threshold-crossing problem in a three-dimensional state space, not a receptor problem. I chose computational neuroscience because it provides the mathematical language to describe how pharmacological interventions alter neural dynamics at the circuit level, not just the molecular level. The ODE/RK45 framework I used for the CCT model is standard in computational neuroscience but absent from pharmacy curricula. The Bayesian MCMC methods I applied to estimate population thresholds are tools I learned from reading Gershman and Daw papers on arXiv, not from any course. This fellowship would replace ad-hoc self-study with structured training, accelerating my ability to contribute to the field. Nigeria has 200 million people, a growing substance-use crisis, and exactly zero computational neuroscience PhD programmes. My long-term goal is to establish a computational addiction neuroscience research group in Lagos that trains African scientists in the methods I am seeking to learn. The fellowship is the first step on that path. SHORT ESSAY: TECHNICAL BACKGROUND Python is my primary research language. I use scipy and numpy for ODE integration, PyMC for Bayesian MCMC sampling, and pandas for data manipulation. The CCT model's ODE system runs on RK45 with adaptive step sizing, producing the encoding probability trajectories that confirmed all five pre-registered hypotheses. The Bayesian population dynamics model uses Hamiltonian Monte Carlo with 4 chains of 2000 samples each, achieving R-hat values below 1.01 for all parameters. For neural simulation, I have working knowledge of NEURON and Brian2. The TOPOLOGIX platform uses Ripser for persistent homology computation and Gudhi for simplicial complex construction, applied to drug-protein interaction networks. The hERG cardiotoxicity MVP within TOPOLOGIX demonstrates my ability to combine topological data analysis with pharmacological safety prediction. My bioinformatics pipeline work with GHRU-GSAR uses Nextflow for workflow management and SLURM for HPC job scheduling, skills directly transferable to large-scale neural simulation. R is used for statistical analysis and visualization. JavaScript and Node.js were used to build the IMPRINT screening platform frontend. Supabase and Postgres handle the backend data storage. This full-stack capability means I can build the computational tools I need rather than waiting for institutional support. CHECKLIST - [ ] Motivation letter, 300-500 words, tailored to the 6 PhD Computational Neuroscience Fellowship - [ ] Research statement, 400-600 words, describing CCT model extension plan - [ ] Short essay on motivation for computational neuroscience, 200-350 words - [ ] Short essay on technical background, 200-350 words - [ ] CV or resume listing degrees, publications, platforms, employment, and skills - [ ] Three letters of recommendation from Kent Berridge, Samuel Gershman, and Nathaniel Daw or Marcelo Mattar - [ ] Transcript from University of Ibadan showing B.Pharm with CGPA 5.1/7.0 - [ ] ORCID profile link: 0009-0001-9272-6735 - [ ] GitHub profile link: github.com/AmunRaPtah - [ ] Links to three sole-authored preprints on OSF and Zenodo - [ ] Link to review article under review at Neuroscience and Biobehavioral Reviews - [ ] Link to co-authored paper under review at Alcohol - [ ] Provisional patent documentation for CCT core architecture - [ ] Proof of PCN pharmacist licensure - [ ] Proof of independent researcher status (no current MSc or PhD enrollment) EDITOR NOTES - Eligibility risk: the programme listing on scholarshipdb.net is generic and may not accept independent researchers without current PhD enrollment. Verify directly with the programme provider whether pre-PhD applicants are eligible. - The programme URL points to a search results page, not a specific fellowship page. The applicant must locate the exact programme page and confirm all requirements, deadlines, and funding amounts before submitting. - The four named endorsers (Berridge, Gershman, Daw, Mattar) have not yet been confirmed as willing to write letters. The applicant must secure written confirmation from at least three before the deadline. - The German equivalent grade of 1.9 is self-calculated. The applicant should verify this conversion with the programme or with a credential evaluation service, as different systems use different scales. - The provisional patent filing date of Q3 2026 may not yet be complete. Confirm the filing status and include the application number if available.