Draft — FRIAS Early Career Fellowship 2026

MOTIVATION LETTER

The Conjunctive Consolidation Threshold model addresses a question that pharmacology has left open for decades: why do some reward memories become permanently encoded while others fade? My work formalizes this as a tripartite dynamical system, coupling dopaminergic reward prediction error, NMDAR-dependent long-term potentiation, and affective contrast into a single ODE framework. The model was calibrated with Bayesian MCMC using PyMC's DEMetropolisZ sampler, with 14 free parameters and literature-elicited priors drawn from a systematic screen of 1,847 records. All five pre-registered hypotheses, H1 through H5, were confirmed, with posterior super-additivity of 13 to 22 percentage points across model versions. Three sole-authored preprints are currently under review at peer-reviewed journals: International Addiction Review, Progress in Neuro-Psychopharmacology and Biological Psychiatry, and Neuroscience and Biobehavioral Reviews.

The FRIAS Early Career Fellowship offers a specific resource that my independent research trajectory lacks: sustained institutional embedding. Since completing my B.Pharm at the University of Ibadan in 2021, I have built and validated this research program without a home laboratory, relying on open-source infrastructure and remote collaboration. The University of Freiburg's strengths in computational neuroscience and systems pharmacology provide the exact environment where the CCT model can move from theoretical framework to experimentally testable predictions. My enrollment in the M.Sc. Digital Health program at Hasso Plattner Institute, University of Potsdam, beginning Winter Semester 2026/27, demonstrates my commitment to the German academic system, and a FRIAS fellowship would allow me to deepen that integration.

My trajectory also speaks to the fellowship's international mobility criterion. I completed my first degree in Nigeria and have since built collaborative relationships with researchers at the University of Michigan, Harvard University, Princeton University, and New York University. Samuel Gershman at Harvard endorsed my arXiv submission; Kent Berridge at Michigan has engaged with the CCT framework's affective contrast axis. The fellowship's 4 to 10 month residency requirement fits my current stage: I can structure the residency around my MSc coursework, which is largely asynchronous and project-based.

The CCT model is a calibrated, pre-registered, and hypothesis-confirmed framework with three manuscripts in peer review. What it lacks is the collaborative infrastructure to design the experimental validations that would move it from computational model to clinical intervention target. Freiburg's neuroscience community, particularly groups working on opioid systems and reward circuitry, offers that infrastructure. I am applying to FRIAS because the fellowship's structure, full-time presence and interdisciplinary mandate, matches what this project needs at this exact stage.

RESEARCH STATEMENT

Project Title: From Computational Model to Experimental Design: Validating the Conjunctive Consolidation Threshold Framework for Addiction Memory Prevention

The Conjunctive Consolidation Threshold model proposes that reward-memory encoding requires simultaneous activation of three coupled axes: dopaminergic reward prediction error, NMDAR-dependent long-term potentiation, and affective contrast between the drug state and the pre-drug baseline. The model is implemented as a system of ordinary differential equations solved with RK45 integration. Bayesian calibration using PyMC's DEMetropolisZ sampler produced posterior distributions for all 14 free parameters, with priors elicited from a structured literature screen of 1,847 records spanning addiction neuroscience, synaptic plasticity, and affective neuroscience. The model's central prediction, that sub-threshold activation on any single axis prevents consolidation, was confirmed across all five pre-registered hypotheses. The super-additivity effect, where combined sub-threshold inputs produce supra-linear consolidation prevention, ranged from 13 to 22 percentage points across model versions.

The project proposed for the FRIAS fellowship has three aims. First, I will extend the CCT model to incorporate receptor-level pharmacokinetic dynamics, linking specific pharmacological agents to their predicted consolidation-prevention profiles. This builds directly on my neurocascade simulation engine, which couples pharmacokinetics to receptor binding to Wilson-Cowan circuit dynamics to behavioral readouts. The mu-opioid, D2 dopamine, and GABA-A receptor systems are already calibrated in that engine, with 62 of 62 tests passing. Second, I will generate a ranked list of testable pharmacological combinations predicted to prevent reward-memory consolidation, each with explicit dose-response predictions derived from the calibrated model. Third, I will design the experimental protocols, including behavioral paradigms and pharmacological dosing schedules, that would test these predictions in rodent models, in collaboration with Freiburg experimental groups.

The methodological foundation for this work is established. My TOPOLOGIX pipeline, which uses ESM-2 protein language model delta-embeddings combined with Morgan fingerprints and a Random Forest classifier, achieves an AUROC of 0.804 plus or minus 0.025 on the Platinum benchmark of 553 drug-resistance mutations, outperforming structure-based baselines such as mCSM-lig at approximately 0.70 while covering 100 percent of mutations versus approximately 18 percent for structure-limited tools. This demonstrates my capacity to build rigorous, reproducible computational pipelines. My cardiotoxicity topology study, which tested whether bipartite persistent homology predicts hERG cardiotoxicity, produced a pre-registered, powered replication showing that topological features do not beat a plain descriptor baseline, AUROC 0.8426 versus 0.8782. That negative result, which settles a comparison the literature had never actually run, exemplifies the methodological rigor I bring to the CCT project.

The fellowship period will focus on the first aim, extending the CCT model with receptor-level dynamics, and producing the ranked pharmacological combination predictions. The experimental protocol designs will be drafted in collaboration with Freiburg experimental neuroscientists, with the goal of establishing a concrete collaboration that can be funded through subsequent grant applications. The full-time presence requirement at FRIAS is compatible with my MSc schedule, as the Digital Health program at HPI is structured around project-based learning with flexible on-site requirements.

The CCT model's potential clinical significance is substantial. Addiction relapse is driven by persistent reward memories that trigger craving upon re-exposure to drug-associated cues. A pharmacological framework that can prevent the initial consolidation of these memories, rather than treating withdrawal symptoms after dependence is established, would represent a fundamental shift in intervention strategy. The model's three-axis structure suggests that combination therapies targeting dopaminergic, glutamatergic, and affective systems simultaneously may achieve consolidation prevention at doses that are individually sub-effective, reducing side-effect burden. The super-additivity effect quantified in my posterior analysis provides the quantitative basis for this prediction.

Freiburg is the right environment for this work. The university's research profile includes strong groups in systems neuroscience, synaptic plasticity, and computational pharmacology. The FRIAS fellowship's interdisciplinary mandate, bringing together researchers from different fields around a shared research question, matches the CCT model's inherent interdisciplinarity. I am not proposing to bring a finished framework to Freiburg; I am proposing to develop the framework's experimental arm in an environment that has the expertise to evaluate and refine it.

EDITOR NOTES

- Eligibility risk: the FRIAS Early Career Fellowship requires a completed doctoral degree. The applicant has a B.Pharm and is enrolled in an M.Sc. program, with no PhD. This is a critical mismatch that must be resolved before submission, either by confirming the fellowship accepts pre-PhD applicants or by identifying an alternative programme.

- The publication requirement, at least 4 peer-reviewed publications since 1 January 2021, is not currently met. The applicant has three sole-authored preprints under review and one co-authored paper under review at Alcohol (Elsevier), but none are yet published. The application should note the under-review status explicitly and may need to be deferred until acceptances are confirmed.

- The international mobility criterion, at least 6 months of research or study in a country different from the country of first academic degree, is met through the HPI/Potsdam enrollment, but the enrollment begins Winter Semester 2026/27. The applicant should confirm whether the enrollment itself satisfies the criterion or whether physical presence in Germany is required before the fellowship start date.

- The applicant must verify whether the fellowship's 4 to 10 month full-time residency requirement can be structured around the HPI MSc schedule. The profile states the program is largely asynchronous, but this needs confirmation from HPI before committing to the fellowship timeline.

- The letter of support requirement applies only to experimental scientists, so it is not needed for this application. However, the applicant should secure at least one letter of recommendation from a named collaborator, such as Kent Berridge or Samuel Gershman, to strengthen the application even if not formally required.

- The project proposal must not exceed 3000 words and 4 graphs. The research statement above is within the word limit but the graphs need to be selected carefully. Recommended figures: the CCT model architecture diagram, the posterior super-additivity plot, the TOPOLOGIX AUROC comparison, and the neurocascade system diagram.

- The CV must use the mandatory FRIAS template. The applicant's tabular CV should emphasize the peer-reviewed publications under review, the ORCID and GitHub identifiers, and the international collaboration network.

- The publication list is capped at 8 entries, with at least 4 peer-reviewed after 2021. The applicant should list the three sole-authored preprints, the co-authored Alcohol paper, and any other publications or preprints from the GHRU-GSAR and CDDDP research assistant roles.

- The layman's abstract must be written in accessible language, avoiding all technical jargon. The applicant should describe the CCT model as a framework for understanding why some memories become permanent and whether that process can be interrupted with medication combinations.

- The motivation statement should address the applicant's specific reasons for choosing Freiburg over other potential host institutions, referencing named research groups or faculty where possible. The current draft mentions Freiburg's strengths generally but should be made more specific after the applicant researches the relevant groups.

- The ethical issue table must address the use of published literature data for model calibration, the absence of human or animal subjects in the current phase, and the ethical considerations of the proposed experimental protocols for future rodent studies.

- The applicant should confirm whether the online form requires the abstract and layman's abstract to be entered separately or as part of the project proposal document. The FRIAS call page should be checked for the exact submission structure.

- The rolling deadline means the applicant should submit as soon as the materials are complete, but the eligibility issues above should be resolved first. A submission that is rejected on eligibility grounds cannot be resubmitted to the same call.

MOTIVATION LETTER

The Conjunctive Consolidation Threshold model addresses a question that pharmacology has left open for decades: why do some reward memories become permanently encoded while others fade? My work formalizes this as a tripartite dynamical system, coupling dopaminergic reward prediction error, NMDAR-dependent long-term potentiation, and affective contrast into a single ODE framework. The model was calibrated with Bayesian MCMC using PyMC's DEMetropolisZ sampler, with 14 free parameters and literature-elicited priors drawn from a systematic screen of 1,847 records. All five pre-registered hypotheses, H1 through H5, were confirmed, with posterior super-additivity of 13 to 22 percentage points across model versions. Three sole-authored preprints are currently under review at peer-reviewed journals: International Addiction Review, Progress in Neuro-Psychopharmacology and Biological Psychiatry, and Neuroscience and Biobehavioral Reviews.

The FRIAS Early Career Fellowship offers a specific resource that my independent research trajectory lacks: sustained institutional embedding. Since completing my B.Pharm at the University of Ibadan in 2021, I have built and validated this research program without a home laboratory, relying on open-source infrastructure and remote collaboration. The University of Freiburg's strengths in computational neuroscience and systems pharmacology provide the exact environment where the CCT model can move from theoretical framework to experimentally testable predictions. My enrollment in the M.Sc. Digital Health program at Hasso Plattner Institute, University of Potsdam, beginning Winter Semester 2026/27, demonstrates my commitment to the German academic system, and a FRIAS fellowship would allow me to deepen that integration.

My trajectory also speaks to the fellowship's international mobility criterion. I completed my first degree in Nigeria and have since built collaborative relationships with researchers at the University of Michigan, Harvard University, Princeton University, and New York University. Samuel Gershman at Harvard endorsed my arXiv submission; Kent Berridge at Michigan has engaged with the CCT framework's affective contrast axis. The fellowship's 4 to 10 month residency requirement fits my current stage: I can structure the residency around my MSc coursework, which is largely asynchronous and project-based.

The CCT model is a calibrated, pre-registered, and hypothesis-confirmed framework with three manuscripts in peer review. What it lacks is the collaborative infrastructure to design the experimental validations that would move it from computational model to clinical intervention target. Freiburg's neuroscience community, particularly groups working on opioid systems and reward circuitry, offers that infrastructure. I am applying to FRIAS because the fellowship's structure, full-time presence and interdisciplinary mandate, matches what this project needs at this exact stage.

RESEARCH STATEMENT

Project Title: From Computational Model to Experimental Design: Validating the Conjunctive Consolidation Threshold Framework for Addiction Memory Prevention

The Conjunctive Consolidation Threshold model proposes that reward-memory encoding requires simultaneous activation of three coupled axes: dopaminergic reward prediction error, NMDAR-dependent long-term potentiation, and affective contrast between the drug state and the pre-drug baseline. The model is implemented as a system of ordinary differential equations solved with RK45 integration. Bayesian calibration using PyMC's DEMetropolisZ sampler produced posterior distributions for all 14 free parameters, with priors elicited from a structured literature screen of 1,847 records spanning addiction neuroscience, synaptic plasticity, and affective neuroscience. The model's central prediction, that sub-threshold activation on any single axis prevents consolidation, was confirmed across all five pre-registered hypotheses. The super-additivity effect, where combined sub-threshold inputs produce supra-linear consolidation prevention, ranged from 13 to 22 percentage points across model versions.

The project proposed for the FRIAS fellowship has three aims. First, I will extend the CCT model to incorporate receptor-level pharmacokinetic dynamics, linking specific pharmacological agents to their predicted consolidation-prevention profiles. This builds directly on my neurocascade simulation engine, which couples pharmacokinetics to receptor binding to Wilson-Cowan circuit dynamics to behavioral readouts. The mu-opioid, D2 dopamine, and GABA-A receptor systems are already calibrated in that engine, with 62 of 62 tests passing. Second, I will generate a ranked list of testable pharmacological combinations predicted to prevent reward-memory consolidation, each with explicit dose-response predictions derived from the calibrated model. Third, I will design the experimental protocols, including behavioral paradigms and pharmacological dosing schedules, that would test these predictions in rodent models, in collaboration with Freiburg experimental groups.

The methodological foundation for this work is established. My TOPOLOGIX pipeline, which uses ESM-2 protein language model delta-embeddings combined with Morgan fingerprints and a Random Forest classifier, achieves an AUROC of 0.804 plus or minus 0.025 on the Platinum benchmark of 553 drug-resistance mutations, outperforming structure-based baselines such as mCSM-lig at approximately 0.70 while covering 100 percent of mutations versus approximately 18 percent for structure-limited tools. This demonstrates my capacity to build rigorous, reproducible computational pipelines. My cardiotoxicity topology study, which tested whether bipartite persistent homology predicts hERG cardiotoxicity, produced a pre-registered, powered replication showing that topological features do not beat a plain descriptor baseline, AUROC 0.8426 versus 0.8782. That negative result, which settles a comparison the literature had never actually run, exemplifies the methodological rigor I bring to the CCT project.

The fellowship period will focus on the first aim, extending the CCT model with receptor-level dynamics, and producing the ranked pharmacological combination predictions. The experimental protocol designs will be drafted in collaboration with Freiburg experimental neuroscientists, with the goal of establishing a concrete collaboration that can be funded through subsequent grant applications. The full-time presence requirement at FRIAS is compatible with my MSc schedule, as the Digital Health program at HPI is structured around project-based learning with flexible on-site requirements.

The CCT model's potential clinical significance is substantial. Addiction relapse is driven by persistent reward memories that trigger craving upon re-exposure to drug-associated cues. A pharmacological framework that can prevent the initial consolidation of these memories, rather than treating withdrawal symptoms after dependence is established, would represent a fundamental shift in intervention strategy. The model's three-axis structure suggests that combination therapies targeting dopaminergic, glutamatergic, and affective systems simultaneously may achieve consolidation prevention at doses that are individually sub-effective, reducing side-effect burden. The super-additivity effect quantified in my posterior analysis provides the quantitative basis for this prediction.

Freiburg is the right environment for this work. The university's research profile includes strong groups in systems neuroscience, synaptic plasticity, and computational pharmacology. The FRIAS fellowship's interdisciplinary mandate, bringing together researchers from different fields around a shared research question, matches the CCT model's inherent interdisciplinarity. I am not proposing to bring a finished framework to Freiburg; I am proposing to develop the framework's experimental arm in an environment that has the expertise to evaluate and refine it.

EDITOR NOTES

- Eligibility risk: the FRIAS Early Career Fellowship requires a completed doctoral degree. The applicant has a B.Pharm and is enrolled in an M.Sc. program, with no PhD. This is a critical mismatch that must be resolved before submission, either by confirming the fellowship accepts pre-PhD applicants or by identifying an alternative programme.

- The publication requirement, at least 4 peer-reviewed publications since 1 January 2021, is not currently met. The applicant has three sole-authored preprints under review and one co-authored paper under review at Alcohol (Elsevier), but none are yet published. The application should note the under-review status explicitly and may need to be deferred until acceptances are confirmed.

- The international mobility criterion, at least 6 months of research or study in a country different from the country of first academic degree, is met through the HPI/Potsdam enrollment, but the enrollment begins Winter Semester 2026/27. The applicant should confirm whether the enrollment itself satisfies the criterion or whether physical presence in Germany is required before the fellowship start date.

- The applicant must verify whether the fellowship's 4 to 10 month full-time residency requirement can be structured around the HPI MSc schedule. The profile states the program is largely asynchronous, but this needs confirmation from HPI before committing to the fellowship timeline.

- The letter of support requirement applies only to experimental scientists, so it is not needed for this application. However, the applicant should secure at least one letter of recommendation from a named collaborator, such as Kent Berridge or Samuel Gershman, to strengthen the application even if not formally required.

- The project proposal must not exceed 3000 words and 4 graphs. The research statement above is within the word limit but the graphs need to be selected carefully. Recommended figures: the CCT model architecture diagram, the posterior super-additivity plot, the TOPOLOGIX AUROC comparison, and the neurocascade system diagram.

- The CV must use the mandatory FRIAS template. The applicant's tabular CV should emphasize the peer-reviewed publications under review, the ORCID and GitHub identifiers, and the international collaboration network.

- The publication list is capped at 8 entries, with at least 4 peer-reviewed after 2021. The applicant should list the three sole-authored preprints, the co-authored Alcohol paper, and any other publications or preprints from the GHRU-GSAR and CDDDP research assistant roles.

- The layman's abstract must be written in accessible language, avoiding all technical jargon. The applicant should describe the CCT model as a framework for understanding why some memories become permanent and whether that process can be interrupted with medication combinations.

- The motivation statement should address the applicant's specific reasons for choosing Freiburg over other potential host institutions, referencing named research groups or faculty where possible. The current draft mentions Freiburg's strengths generally but should be made more specific after the applicant researches the relevant groups.

- The ethical issue table must address the use of published literature data for model calibration, the absence of human or animal subjects in the current phase, and the ethical considerations of the proposed experimental protocols for future rodent studies.

- The applicant should confirm whether the online form requires the abstract and layman's abstract to be entered separately or as part of the project proposal document. The FRIAS call page should be checked for the exact submission structure.

- The rolling deadline means the applicant should submit as soon as the materials are complete, but the eligibility issues above should be resolved first. A submission that is rejected on eligibility grounds cannot be resubmitted to the same call.

AI Draft — FRIAS Early Career Fellowship 2026