MOTIVATION LETTER The CCT model, a tripartite pharmacological framework for reward-memory encoding prevention in addiction, was built from a pharmacist's understanding of receptor dynamics and a computational modeler's toolkit. Five pre-registered hypotheses, tested with a 14-parameter ODE model calibrated via Bayesian MCMC against 1,847 literature records, all confirmed. Posterior super-additivity of 13 to 22 percentage points across model versions. This is the kind of cross-domain computational neuroscience the programme at scholarshipdb.net seeks to support: mechanistic, falsifiable, and grounded in pharmacology. I am a 29-year-old Nigerian pharmacist and independent computational researcher. My B.Pharm from the University of Ibadan (CGPA 5.1/7.0, German equivalent 1.9) gave me the pharmacology foundation. My research since 2024 has produced three sole-authored preprints on the CCT model, a co-authored paper under review at Alcohol (Elsevier), and four additional pre-registered studies spanning hERG cardiotoxicity topology, drug-resistance prediction, brain-circuit simulation, and macromolecular transport modeling. I am enrolled in the M.Sc. Digital Health at Hasso Plattner Institute / University of Potsdam starting Winter Semester 2026/27, which provides a direct institutional bridge to German computational neuroscience. The programme's selection criteria include academic excellence, novelty of research proposal, relevance to computational neuroscience, interdisciplinary potential, and career stage fit. My CCT model directly addresses all five. It integrates dynamical systems (ODE/RK45), Bayesian inference (PyMC DEMetropolisZ), and pharmacological mechanism (dopaminergic RPE, NMDAR-dependent LTP, affective contrast). It is novel: no existing model treats reward-memory consolidation as a conjunctive threshold across three coupled axes. It is relevant: addiction neuroscience is a core computational neuroscience domain. It is interdisciplinary: pharmacy, machine learning, and dynamical systems. And I am early-career, pre-PhD, with endorsements from Kent Berridge (Michigan), Samuel Gershman (Harvard), Nathaniel Daw (Princeton), and Marcelo Mattar (NYU). The programme supports independent researchers. I have operated independently since 2024, managing my own pre-registration, data collection, analysis, and preprint pipeline. My hERG topology study was a pre-registered, powered replication that settled a comparison the published literature had never actually run: topological features do not beat a plain descriptor baseline (AUROC 0.8426 vs 0.8782). My drug-resistance prediction work (TOPOLOGIX) uses ESM-2 protein-language-model delta-embeddings plus Morgan fingerprints and a Random Forest classifier, achieving AUROC 0.804 on the Platinum benchmark, beating structure-based baselines while covering 100% of mutations versus approximately 18% for structure-limited tools. I seek this programme to formalize my training in computational neuroscience methods, connect with German research groups working on reinforcement learning and addiction, and develop the CCT model into a full PhD project. The programme's support would enable me to attend workshops, access computational resources, and establish collaborations that my independent status currently limits. RESEARCH STATEMENT The central problem I address is how pharmacological interventions can prevent the encoding of reward-associated memories that drive addiction relapse. Existing computational models of addiction focus on either dopaminergic reward prediction error or synaptic plasticity, but none integrate the three axes that my CCT model specifies: dopaminergic RPE signaling, NMDAR-dependent long-term potentiation, and affective contrast between drug and natural rewards. My hypothesis is that a conjunctive consolidation threshold exists across these three axes, and that crossing this threshold is necessary and sufficient for reward-memory encoding. I tested this hypothesis with a system of three coupled ordinary differential equations solved via RK45 integration. The model has 14 free parameters, each with literature-elicited priors from a systematic screen of 1,847 records. I calibrated the model using Bayesian MCMC (PyMC DEMetropolisZ) against published behavioral and pharmacological data. All five pre-registered hypotheses (H1-H5) were confirmed. Posterior super-additivity ranged from 13 to 22 percentage points across model versions, meaning the combined effect of the three axes exceeds the sum of individual effects. This supports the conjunctive threshold framework. The CCT model has direct implications for medication development. It predicts that interventions targeting any single axis will be insufficient; effective prevention requires simultaneous modulation of at least two axes. For example, a D2 antagonist alone reduces RPE but leaves NMDAR plasticity and affective contrast intact, so the threshold can still be crossed. A combination of a D2 antagonist and an NMDAR partial agonist reduces two axes simultaneously, dropping the system below threshold. This prediction is testable in rodent self-administration paradigms. My broader research programme extends this framework in three directions. First, I am developing neurocascade, a receptor-to-behavior brain-circuit simulation engine that couples pharmacokinetics to receptor binding to Wilson-Cowan circuit dynamics to behavioral readout. Three literature-calibrated receptor/circuit systems (mu-opioid, D2 dopamine, GABA-A) have been implemented, with 62 of 62 tests passing. Second, I am applying topological data analysis to protein-ligand interfaces, having shown that bipartite persistent homology does not predict hERG cardiotoxicity beyond a plain descriptor baseline (AUROC 0.8426 vs 0.8782) and does not predict drug resistance (AUROC 0.425 and 0.485 on the Platinum benchmark). These negative results rule out interface geometry as the driver and motivate sequence-representation approaches like my TOPOLOGIX pipeline (AUROC 0.804 on Platinum). Third, I am extending Koopman operator methods with Mori-Zwanzig memory kernels for modeling drug transport through tumor tissue, with a pre-registered gated validation pipeline that transparently reports failures. The programme at scholarshipdb.net supports computational neuroscience that combines methods, models, and data analysis. My work does exactly this: ODE modeling, Bayesian inference, topological data analysis, protein language models, and dynamical systems theory, all applied to concrete pharmacological and neuroscientific questions. The programme's emphasis on interdisciplinary potential matches my background as a pharmacist-computational modeler-software engineer. My upcoming MSc at HPI/Potsdam provides institutional continuity for a longer-term research trajectory in Germany. SHORT ESSAY: RELEVANCE TO COMPUTATIONAL NEUROSCIENCE Computational neuroscience requires models that bridge levels of analysis: from molecules to circuits to behavior. My CCT model does this explicitly. The dopaminergic RPE axis connects to temporal difference learning theory. The NMDAR-dependent LTP axis connects to synaptic plasticity models. The affective contrast axis connects to opponent-process theory. The ODE framework integrates these into a single dynamical system that makes testable predictions about pharmacological intervention timing and combination. My neurocascade engine extends this bridging further. It couples pharmacokinetic compartment models to receptor binding kinetics to Wilson-Cowan neural population dynamics to behavioral readout. This allows simulation of how a drug dose, administered at a specific time, propagates through the body, binds to receptors, alters circuit dynamics, and changes behavior. The engine is modular: any receptor system, any circuit architecture, any behavioral readout can be plugged in. This is computational neuroscience as engineering: building tools that let researchers test hypotheses about multi-level mechanisms. My topological data analysis work, while producing negative results, is equally relevant. Testing whether persistent homology of protein-ligand interfaces predicts functional outcomes is a computational neuroscience question: it asks whether geometric features of molecular interactions carry information about neural system responses. The answer, for hERG cardiotoxicity and drug resistance, is no. This negative result constrains future model building and demonstrates methodological rigor. SHORT ESSAY: INTERDISCIPLINARY POTENTIAL AND SKILL SET I hold a pharmacy degree, build ODE models in Python, calibrate them with Bayesian MCMC, analyze protein structures with AlphaFold and RDKit, simulate neural circuits with NEURON and Brian2, and deploy production data pipelines with DuckDB and Supabase. This combination is unusual. Most computational neuroscientists come from physics, computer science, or neuroscience; few bring a pharmacist's detailed knowledge of receptor pharmacology, drug metabolism, and clinical dosing. This interdisciplinary background gives me a specific advantage: I can build models that are pharmacologically realistic from the start. The CCT model's parameters are not arbitrary; they are constrained by literature on D2 receptor occupancy, NMDAR antagonist dosing, and dopamine reuptake kinetics. My hERG topology study used actual protein-ligand structures from the PDB, not idealized representations. My TOPOLOGIX pipeline uses real mutation data from the Platinum benchmark. The programme supports interdisciplinary researchers. My skill set spans computational methods (Python, R, ODE/RK45, PyMC, TDA, protein language models), wet-lab adjacent skills (pharmacology, ADMET/QSAR, molecular docking), and infrastructure (Nextflow/SLURM/HPC, self-hosted LLM serving, CI/CD). I am equally comfortable discussing dopamine receptor subtypes and debugging a PyMC sampling chain. SHORT ESSAY: CAREER STAGE AND FIT FOR EARLY-CAREER TRACK I am 29, pre-PhD, with an independent research record of five pre-registered studies, three sole-authored preprints, one co-authored paper under review, and endorsements from four established researchers. I have no institutional research position. My research is self-funded, self-managed, and self-archived on OSF and Zenodo. This is the definition of an early-career independent researcher. The programme's support would formalize my training. I am enrolled in the M.Sc. Digital Health at HPI/Potsdam starting Winter 2026/27, which provides a structured academic environment. The programme would supplement this with computational neuroscience-specific training, workshops, and networking. My goal is to develop the CCT model into a PhD project at a German university, with potential supervisors including those in the Berlin computational neuroscience community. My Nigeria background adds a unique perspective. Addiction research in Africa is underdeveloped, with few computational models and limited pharmacological data. My work can inform medication development for populations with different genetic backgrounds, drug use patterns, and healthcare systems. The programme's support for LMIC-track researchers aligns with this. CHECKLIST - [ ] Motivation letter (300-500 words, written above) - [ ] Research statement (400-600 words, written above) - [ ] Short essay: Relevance to computational neuroscience (200-350 words, written above) - [ ] Short essay: Interdisciplinary potential and skill set (200-350 words, written above) - [ ] Short essay: Career stage and fit for early-career track (200-350 words, written above) - [ ] CV or resume (applicant must prepare from profile data) - [ ] Academic transcripts (B.Pharm from University of Ibadan, proof of MSc enrollment at HPI/Potsdam) - [ ] Proof of ORCID (0009-0001-9272-6735) and GitHub (github.com/AmunRaPtah) - [ ] Letters of recommendation or endorsements (contact Kent Berridge, Samuel Gershman, Nathaniel Daw, Marcelo Mattar) - [ ] Preprint links (OSF/Zenodo for CCT model, hERG topology, drug resistance, neurocascade, ergofluids) - [ ] Co-authored paper under review at Alcohol (Elsevier) - provide manuscript or acceptance letter if available - [ ] Proof of PCN pharmacist license - [ ] Proof of Nigerian nationality (passport or national ID) - [ ] Programme application form at scholarshipdb.net (complete online) EDITOR NOTES - Eligibility risk: The programme URL points to a general scholarship database page, not a specific programme. The provider, amount, and deadline are unknown. Verify the exact programme name, eligibility criteria, and application portal before submitting. If this is a database of scholarships, the applicant must identify a specific programme from the list and tailor materials accordingly. - Facts needing verification: The German equivalent GPA of 1.9 for a 5.1/7.0 CGPA should be confirmed with the University of Ibadan's grading scale and the standard conversion used by German universities. Some programmes require a certified translation. - Gap: The applicant's employment history (National Product Manager at Synthcare, Clinical Pharmacist at Ramset Pharmacy) is not directly relevant to computational neuroscience. The motivation letter and research statement should emphasize the research work, not the employment. However, the applicant should prepare a brief explanation of how these roles connect to their research interests if asked. - Gap: The applicant's MSc at HPI/Potsdam starts Winter Semester 2026/27. If the programme deadline is before this date, the applicant should clarify their enrollment status (accepted, conditional, or pending) and provide proof of admission or application. - Gap: The applicant's endorsements from Berridge, Gershman, Daw, and Mattar are listed but not described. The applicant should confirm that these researchers are willing to provide letters or endorsements, and specify the nature of the relationship (e.g., correspondence about the CCT model, arXiv endorsement, collaboration).