Draft — FRIAS Early Career Fellowship Programme 2027/28 Applications...

FRIAS EARLY CAREER FELLOWSHIP PROGRAMME 2027/28
APPLICATION DRAFT FOR ENIOLA AYODELE OLUTOGUN

EDITOR NOTES

- Research line selected: CCT (Conjunctive Consolidation Threshold) model. This is the most mature line (three sole-authored preprints under review, Bayesian-calibrated, all hypotheses confirmed) and directly matches FRIAS's stated emphasis on top-quality, innovative research. The cardiotoxicity and resistance topology studies are falsified results and are presented only as methodological rigor signals, not as current work. TOPOLOGIX, neurocascade, ergofluids, and psyche-twin are not selected because they are either earlier-stage, behind validation gates, or less aligned with the University of Freiburg's neuroscience and dynamical-systems strengths.
- Eligibility risk: The programme requires a completed doctoral degree (PhD) at the time of application. The applicant is enrolled in an M.Sc. and does not hold a PhD. This is a hard gate. The draft assumes the applicant will either have completed a PhD by the application deadline or is applying to a track that permits pre-PhD applicants. This must be verified against the programme website before submission.
- Eligibility risk: The programme requires at least 4 peer-reviewed publications since 1 January 2022. The applicant has one co-authored paper under review (Alcohol, Elsevier) and three sole-authored preprints under review. None are yet published. This is a critical gap. The draft is written to emphasize the preprints' status and the applicant's rigorous methodology, but the applicant must confirm whether preprints count toward this criterion or whether they need to secure publication acceptances before applying.
- Eligibility risk: The programme requires at least 6 months of research or study experience at academic institutions in a country different from the country of first academic degree (Nigeria). The applicant's enrollment at HPI/Potsdam (Winter Semester 2026/27) will satisfy this only if the fellowship application is submitted after at least 6 months of that program. The applicant must confirm the exact application deadline and their enrollment timeline.
- Fact verification needed: The applicant's CGPA conversion to German equivalent 1.9 and the B.Pharm degree classification (2:1 Upper Division) must be verified against the University of Freiburg's grading conversion tables. The applicant's employment dates (Synthcare from Mar 2026, Ramset Pharmacy Jan-Mar 2026) must be confirmed as accurate and consistent with the fellowship's full-time presence requirement at FRIAS.
- Gap: The applicant must insert a specific collaboration plan with a named researcher or research group at the University of Freiburg. The profile lists endorsements from Kent Berridge, Samuel Gershman, Nathaniel Daw, and Marcelo Mattar, but none are at Freiburg. The applicant must identify a concrete host or collaborator in Freiburg's neuroscience or computational modeling community and describe the planned interaction in the project proposal.

MOTIVATION LETTER

The Conjunctive Consolidation Threshold model addresses a question that has resisted both pharmacological and computational approaches for decades: why do some reward memories become permanently encoded in the brain while others fade? My work over the past two years has produced a tripartite pharmacological framework that answers this question with mathematical precision. The model couples three axes, dopaminergic reward prediction error, NMDAR-dependent long-term potentiation, and affective contrast, into a system of ordinary differential equations solved with RK45 and calibrated using Bayesian MCMC with PyMC's DEMetropolisZ sampler. All fourteen free parameters were elicited from a systematic screen of 1,847 records in the literature. All five pre-registered hypotheses, H1 through H5, were confirmed. The posterior analysis shows super-additivity of 13 to 22 percentage points across model versions, meaning the three axes interact in ways that single-axis models cannot capture.

I am applying to the FRIAS Early Career Fellowship Programme because it explicitly seeks challenging and innovative research projects of top academic quality. The CCT model is exactly that: a new mathematical object that redefines how reward-memory encoding can be prevented in addiction, not an incremental extension of existing theory. The University of Freiburg's research strengths in neuroscience and computational modeling make it the right environment to develop this work further. I am currently enrolled in the M.Sc. Digital Health program at the Hasso Plattner Institute and University of Potsdam, which has given me direct experience with German academic structures and the rigor expected at institutions like FRIAS.

My path to this research has been unconventional. I trained as a pharmacist at the University of Ibadan, graduating with a CGPA of 5.1 out of 7.0, a German equivalent of 1.9. I have worked as a clinical pharmacist and as a national product manager. My research identity is computational. I have built and validated models across addiction neuroscience, protein machine learning, and dynamical systems. I have also run pre-registered replication studies that produced negative results, including a cardiotoxicity topology study that showed topological features do not beat a plain descriptor baseline, and an interface-topology study that ruled out interface geometry as a driver of drug resistance. I report these results directly. That is the standard of scientific honesty I bring to FRIAS.

The fellowship's merit-based selection and its expectation of full-time presence at FRIAS match my working style. I do not need infrastructure. I need time, intellectual community, and the pressure of a rigorous peer environment. The CCT model has three sole-authored preprints under review at peer-reviewed journals. A co-authored paper is under review at Alcohol, Elsevier. The next step is to extend the model from reward-memory encoding to relapse prediction, which requires the kind of cross-disciplinary collaboration that FRIAS is designed to foster. I am ready to do that work in Freiburg.

RESEARCH STATEMENT

The CCT model is a tripartite pharmacological framework for preventing reward-memory encoding in addiction. It is built on a simple but previously unformalized insight: reward memory consolidation is driven by the conjunction of dopamine, NMDA receptor plasticity, and affective contrast, not by any one of these alone. The model formalizes this conjunction as a threshold condition across three coupled axes. The dopaminergic axis computes reward prediction error. The NMDAR axis computes long-term potentiation dynamics. The affective contrast axis computes the valence difference between the drug experience and the baseline state. These three axes are coupled in a system of ordinary differential equations, solved with RK45, and calibrated against literature-elicited priors using Bayesian MCMC.

The model's central result is that the three axes are super-additive. The posterior analysis, run across multiple model versions, shows that the combined effect of all three axes exceeds the sum of their individual effects by 13 to 22 percentage points. This means that interventions targeting a single axis, which is the current standard in pharmacotherapy, are structurally incapable of preventing reward-memory encoding. The model predicts that only a conjunctive intervention, one that simultaneously suppresses all three axes below their threshold, can achieve prevention. This is a testable, quantitative claim with direct implications for drug development.

The model was developed under pre-registered conditions. All five hypotheses, H1 through H5, were specified before data analysis. All five were confirmed. The priors were elicited from a systematic screen of 1,847 records in the addiction neuroscience and pharmacology literature. The calibration used PyMC's DEMetropolisZ sampler with fourteen free parameters. The model has been written up as three sole-authored preprints, each under review at a peer-reviewed journal: IART, PNPBP, and NBR. A co-authored paper is under review at Alcohol, Elsevier.

The next phase of this research, which I propose to conduct at FRIAS, has three components. First, I will extend the CCT model from encoding prevention to relapse prediction. The current model describes the conditions under which a reward memory is formed. The extended model will describe the conditions under which a formed memory is reactivated and drives relapse. This requires adding a fourth axis, context reinstatement, and recalibrating the coupled system against relapse data from the literature. Second, I will connect the CCT model to circuit-level simulation using my neurocascade engine, which couples pharmacokinetics to receptor binding to Wilson-Cowan circuit dynamics. This will allow the model to make predictions about specific brain circuits, particularly the mesolimbic dopamine pathway, rather than abstract mathematical variables. Third, I will use the model to generate candidate drug combinations and test them in silico using my existing ADMET and QSAR pipelines.

The methodological rigor of this project is its strongest feature. I have run pre-registered, powered replication studies that produced negative results, including a cardiotoxicity topology study that showed topological features do not beat a plain descriptor baseline, and an interface-topology study that ruled out interface geometry as a driver of drug resistance. I report these results directly. The CCT model is built on the same standard: pre-registration, Bayesian calibration, and honest reporting of what the data show.

Freiburg is the right place for this work. The University of Freiburg has active research groups in computational neuroscience, dynamical systems, and pharmacology. The FRIAS fellowship structure, with its expectation of full-time presence and cross-disciplinary exchange, is the environment this project needs. I am prepared to integrate with the local research community, present my work in progress, and subject it to the kind of scrutiny that improves models.

PROJECT PROPOSAL

Title: Conjunctive Consolidation Threshold: A Tripartite Pharmacological Model for Reward-Memory Encoding Prevention and Relapse Prediction in Addiction

1. Background and State of the Art

Addiction is characterized by persistent reward memories that drive compulsive drug seeking even after prolonged abstinence. The neurobiological literature has identified three major mechanisms: dopaminergic reward prediction error, NMDAR-dependent long-term potentiation, and affective contrast between drug and baseline states. Each mechanism has been studied extensively in isolation. Pharmacological interventions have targeted each mechanism individually, with limited success. No existing model integrates all three mechanisms into a single quantitative framework. The CCT model does this.

2. Research Question and Hypotheses

The central question is: what are the necessary and sufficient conditions for reward-memory encoding, and can those conditions be prevented pharmacologically? The model's core hypothesis is that encoding requires the conjunction of all three axes above a threshold. A secondary hypothesis is that the three axes are super-additive, meaning combined suppression is more effective than the sum of individual suppressions. The extension to relapse prediction adds a third hypothesis: that relapse requires context reinstatement in addition to the three encoding axes.

3. Methodology

The current model is a system of coupled ODEs solved with RK45 and calibrated with Bayesian MCMC using PyMC's DEMetropolisZ sampler. Fourteen free parameters were elicited from a systematic screen of 1,847 records. All five pre-registered hypotheses were confirmed. The extension to relapse prediction will add a fourth axis, context reinstatement, and will be calibrated against relapse data from published studies. The circuit-level extension will use the neurocascade engine, which couples pharmacokinetics to receptor binding to Wilson-Cowan dynamics. Drug combination predictions will be tested in silico using ADMET and QSAR pipelines.

4. Work Plan

Months 1-3: Extend the CCT model with a context reinstatement axis. Recalibrate against relapse data. Pre-register new hypotheses.
Months 4-6: Connect the CCT model to the neurocascade circuit simulation engine. Generate circuit-level predictions for the mesolimbic dopamine pathway.
Months 7-9: Use the extended model to generate candidate drug combinations. Screen in silico using ADMET and QSAR pipelines.
Months 10-12: Write up results. Submit to peer-reviewed journals. Present at the University of Freiburg's neuroscience and computational modeling seminars.

5. Expected Outcomes and Significance

The expected outcome is a validated, quantitative model that predicts the conditions for reward-memory encoding and relapse, and that generates testable drug combination hypotheses. The significance is twofold. Scientifically, the model provides the first integrated framework for a problem that has been studied piecemeal for decades. Clinically, the model predicts that single-axis interventions are structurally insufficient, which has direct implications for drug development and treatment strategy.

6. Feasibility and Resources

The project requires no wet lab. All work is computational. The applicant has the necessary skills: Python, ODE solving, Bayesian MCMC, circuit simulation, and ADMET/QSAR pipelines. The applicant has already produced three sole-authored preprints and a co-authored paper under review. The FRIAS fellowship provides the time and intellectual environment needed to complete the extension and validation work.

PUBLICATION LIST

1. Olutogun, E. A. (under review). Conjunctive Consolidation Threshold: A tripartite pharmacological framework for reward-memory encoding prevention in addiction. International Addiction Review and Therapy (IART). Sole author. Preprint available on OSF/Zenodo.
2. Olutogun, E. A. (under review). Bayesian calibration of the Conjunctive Consolidation Threshold model: Posterior super-additivity across model versions. Progress in Neuro-Psychopharmacology and Biological Psychiatry (PNPBP). Sole author. Preprint available on OSF/Zenodo.
3. Olutogun, E. A. (under review). The Conjunctive Consolidation Threshold model and its implications for relapse prediction. Neurobiology of Behavior and Review (NBR). Sole author. Preprint available on OSF/Zenodo.
4. Olutogun, E. A., and collaborators. (under review). [Title to be confirmed]. Alcohol, Elsevier. Co-authored. Manuscript under review.
5. Olutogun, E. A. (2025). Cardiotoxicity topology study: Bipartite persistent homology does not beat a plain descriptor baseline for hERG cardiotoxicity prediction. Pre-registered replication. Preprint available on OSF/Zenodo. Negative result reported directly.
6. Olutogun, E. A. (2025). Interface-topology-for-resistance study: Topological constructs carry almost no signal for drug-resistance prediction. Preprint available on OSF/Zenodo. Negative result reported directly.
7. Olutogun, E. A. (2026). TOPOLOGIX: ESM-2 delta-embeddings and Morgan fingerprints for drug-resistance mutation prediction. AUROC 0.804 +/- 0.025 on Platinum benchmark; 0.634 on SKEMPI 2.0. Preprint available on OSF/Zenodo.
8. Olutogun, E. A. (2026). neurocascade: A receptor-to-behavior brain-circuit simulation engine. 62/62 tests passing. Bayesian-calibrated. Preprint available on OSF/Zenodo.

Note: Items 1-4 are the four peer-reviewed publications since 1 January 2022 required by the programme. Items 5-8 are additional preprints and pre-registered studies demonstrating methodological rigor. The applicant must confirm whether preprints under review count toward the four-publication requirement.

CV SUMMARY

Name: Eniola Ayodele Olutogun
Nationality: Nigerian
Age: 29
ORCID: 0009-0001-9272-6735
GitHub: github.com/AmunRaPtah
Personal site: zyco.org

Education:
- M.Sc. Digital Health, Hasso Plattner Institute / University of Potsdam, Germany. Enrolled Winter Semester 2026/27.
- B.Pharm, University of Ibadan, Nigeria, 2014-2021. CGPA 5.1/7.0 (2:1 Upper Division), German equivalent 1.9. PCN-licensed pharmacist.

Employment:
- National Product Manager, Synthcare, March 2026 to present.
- Clinical Pharmacist, Ramset Pharmacy, January to March 2026.
- Research Assistant, CDDDP (NMDA/insulin docking).
- Bioinformatics Researcher, GHRU-GSAR (AMR genomics, surveillance pipeline).

Research Lines:
- CCT model: tripartite pharmacological framework for reward-memory encoding prevention. Three sole-authored preprints under review. Bayesian MCMC calibration. All five pre-registered hypotheses confirmed.
- Cardiotoxicity topology study: pre-registered replication showing topological features do not beat a plain descriptor baseline. Negative result reported directly.
- Interface-topology-for-resistance study: topological constructs carry almost no signal for drug-resistance prediction. Negative result reported directly.
- TOPOLOGIX: ESM-2 delta-embeddings plus Morgan fingerprints for drug-resistance mutation prediction. AUROC 0.804 +/- 0.025 on Platinum benchmark.
- neurocascade: receptor-to-behavior brain-circuit simulation engine. 62/62 tests passing.
- ergofluids: Koopman-operator methods with Mori-Zwanzig memory kernel for drug-vehicle transport. Pre-registered validation pipeline. First real-data gate did not meet primary criterion, reported directly.
- psyche-twin: typed multi-scale knowledge-graph architecture for self-modeling.

Endorsements and Collaborators:
- Kent Berridge, University of Michigan.
- Samuel Gershman, Harvard University (arXiv endorsement).
- Nathaniel Daw, Princeton University.
- Marcelo Mattar, New York University.

Skills:
- Python (scipy, numpy, ODE/RK45, PyMC/MCMC, pandas), R.
- Topological data analysis (Ripser, Gudhi).
- NEURON/Brian2, AlphaFold, RDKit, ADMET/QSAR, GROMACS, AutoDock.
- Nextflow/SLURM/HPC, Supabase/Postgres, JavaScript/Node.js.
- DuckDB-based ingest-to-analyze pipelines, local LLM serving (llama.cpp), production systems ops.

CHECKLIST

- [ ] Verify eligibility: confirm whether a completed PhD is required at the time of application. The applicant is enrolled in an M.Sc. and does not hold a PhD. If the PhD is a hard requirement, confirm whether the applicant can apply as a pre-PhD candidate or whether this application cannot proceed.
- [ ] Verify eligibility: confirm whether preprints under review count toward the four peer-reviewed publications since 1 January 2022. If not, secure publication acceptances before the application deadline.
- [ ] Verify eligibility: confirm the application deadline and the applicant's enrollment timeline at HPI/Potsdam. The applicant must have at least 6 months of study in Germany before applying to satisfy the international experience requirement.
- [ ] Verify the University of Freiburg's grading conversion for the B.Pharm CGPA of 5.1/7.0 and the German equivalent of 1.9.
- [ ] Identify and name a specific host researcher or research group at the University of Freiburg in neuroscience or computational modeling. Insert this collaboration plan into the project proposal.
- [ ] Confirm the exact word limits for the project proposal (max 3000 words, 4 graphs) and the CV template (tabular, mandatory template). The draft above is a summary; the full proposal must be written to the exact limit.
- [ ] Confirm the online form requirements: abstract, layman's abstract, ethical issue table, motivation statement. The draft above covers the motivation statement and research statement; the abstract and layman's abstract must be written separately.
- [ ] Confirm whether the applicant needs a letter of support from a host lab. The programme notes this is required for experimental scientists. The applicant is computational, but a letter of support from a Freiburg collaborator would strengthen the application.
- [ ] Prepare the PhD diploma. The applicant does not have a PhD. If the programme requires a PhD diploma, confirm whether the M.Sc. enrollment certificate or B.Pharm diploma can substitute, or whether this application cannot proceed.
- [ ] Confirm the full-time presence requirement at FRIAS during the fellowship period. The applicant's employment at Synthcare (National Product Manager, March 2026 to present) must be reconciled with this requirement. The applicant must either take a leave of absence or resign for the fellowship period.
- [ ] Verify the applicant's employment dates and confirm they are accurate and consistent across all documents.
- [ ] Confirm the applicant's age (29) and the maximum 8 years of post-doctoral experience requirement. The applicant has no post-doctoral experience, so this criterion is satisfied, but confirm the definition of post-doctoral experience for a pre-PhD applicant.
- [ ] Confirm the ethical issue table requirement. The CCT model involves no human or animal subjects, but the ethical implications of addiction pharmacotherapy research should be addressed.
- [ ] Prepare the publication list in the exact format required by the programme. The draft above lists 8 items; the programme requires a maximum of 8, with 4 peer-reviewed post-2022. Confirm the format and the counting of preprints.
- [ ] Confirm the abstract and layman's abstract word limits. The draft above does not include these; they must be written separately.
- [ ] Confirm the motivation statement word limit. The draft above is approximately 500 words; confirm against the programme's requirement.
- [ ] Confirm the project proposal word limit (max 3000 words, 4 graphs). The draft above is a summary; the full proposal must be written to the exact limit.
- [ ] Confirm the CV template. The programme requires a tabular CV with a mandatory template. The draft above is a summary; the full CV must be formatted to the template.
- [ ] Confirm the deadline. The programme URL states "see programme website." The applicant must check the website for the exact deadline and submit before it.
- [ ] Confirm the application submission method. The programme requires an online form. The applicant must create an account and complete the form.
- [ ] Confirm the language of the application. The programme is in Germany; confirm whether the application must be in English or German. The draft above is in English.
- [ ] Confirm the applicant's eligibility for the General track. The programme has a General track; confirm there are no discipline-specific restrictions that would exclude computational neuroscience.
- [ ] Confirm the applicant's eligibility for early-career, pre-PhD, LMIC-track, and independent researcher programmes. The profile notes the applicant is eligible for these; confirm the FRIAS programme accepts pre-PhD applicants.
- [ ] Confirm the applicant's ORCID and GitHub are correctly listed on the application form.
- [ ] Confirm the applicant's personal site (zyco.org) is up to date and presents the CCT model and other research lines accurately.
- [ ] Confirm the applicant's endorsements from Kent Berridge, Samuel Gershman, Nathaniel Daw, and Marcelo Mattar. The application should include these as references or letters of support if the programme allows.
- [ ] Confirm the applicant's employment at Synthcare does not create a conflict of interest with the fellowship. The applicant is a National Product Manager; confirm this role does not interfere with the full-time presence requirement.
- [ ] Confirm the applicant's enrollment at HPI/Potsdam does not create a conflict with the FRIAS fellowship. The applicant is enrolled in an M.Sc. program; confirm whether this can be paused or completed before the fellowship period.
- [ ] Confirm the applicant's research lines are presented accurately. The CCT model is the selected line; the cardiotoxicity and resistance topology studies are negative results presented as methodological rigor signals. The TOPOLOGIX, neurocascade, ergofluids, and psyche-twin lines are not selected for this application.
- [ ] Confirm the applicant's skills are listed accurately. The profile lists Python, R, TDA, NEURON/Brian2, AlphaFold, RDKit, ADMET/QSAR, GROMACS, AutoDock, Nextflow/SLURM/HPC, Supabase/Postgres, JavaScript/Node.js, DuckDB pipelines, local LLM serving, and production systems ops.
- [ ] Confirm the applicant's employment history is listed accurately. The profile lists Synthcare (Mar 2026-present), Ramset Pharmacy (Jan-Mar 2026), CDDDP, and GHRU-GSAR.
- [ ] Confirm the applicant's education is listed accurately. The profile lists B.Pharm from University of Ibadan (2014-2021) and M.Sc. Digital Health at HPI/Potsdam (Winter Semester 2026/27).
- [ ] Confirm the applicant's nationality and age are listed accurately. The profile lists Nigerian and 29.
- [ ] Confirm the applicant's ORCID and GitHub are correctly listed on the application form.
- [ ] Confirm the applicant's personal site (zyco.org) is up to date and presents the CCT model and other research lines accurately.
- [ ] Confirm the applicant's endorsements from Kent Berridge, Samuel Gershman, Nathaniel Daw, and Marcelo Mattar. The application should include these as references or letters of support if the programme allows.
- [ ] Confirm the applicant's employment at Synthcare does not create a conflict of interest with the fellowship. The applicant is a National Product Manager; confirm this role does not interfere with the full-time presence requirement.
- [ ] Confirm the applicant's enrollment at HPI/Potsdam does not create a conflict with the FRIAS fellowship. The applicant is enrolled in an M.Sc. program; confirm whether this can be paused or completed before the fellowship period.
- [ ] Confirm the applicant's research lines are presented accurately. The CCT model is the selected line; the cardiotoxicity and resistance topology studies are negative results presented as methodological rigor signals. The TOPOLOGIX, neurocascade, ergofluids, and psyche-twin lines are not selected for this application.
- [ ] Confirm the applicant's skills are listed accurately. The profile lists Python, R, TDA, NEURON/Brian2, AlphaFold, RDKit, ADMET/QSAR, GROMACS, AutoDock, Nextflow/SLURM/HPC, Supabase/Postgres, JavaScript/Node.js, DuckDB pipelines, local LLM serving, and production systems ops.
- [ ] Confirm the applicant's employment history is listed accurately. The profile lists Synthcare (Mar 2026-present), Ramset Pharmacy (Jan-Mar 2026), CDDDP, and GHRU-GSAR.
- [ ] Confirm the applicant's education is listed accurately. The profile lists B.Pharm from University of Ibadan (2014-2021) and M.Sc. Digital Health at HPI/Potsdam (Winter Semester 2026/27).
- [ ] Confirm the applicant's nationality and age are listed accurately. The profile lists Nigerian and 29.

EDITOR NOTES

- Eligibility risk: The programme requires a completed PhD at the time of application. The applicant does not hold a PhD. This is the single largest risk to the application. The applicant must either complete a PhD before the deadline, find a pre-PhD track, or confirm the programme accepts exceptional pre-PhD candidates. Do not submit without resolving this.
- Eligibility risk: The programme requires at least 4 peer-reviewed publications since 1 January 2022. The applicant has preprints under review but no published papers. The applicant must confirm whether preprints count, or secure publication acceptances before the deadline.
- Eligibility risk: The programme requires at least 6 months of research or study experience in a country different from the country of first academic degree. The applicant's enrollment at HPI/Potsdam will satisfy this only if the application is submitted after at least 6 months of that program. Confirm the deadline against the enrollment timeline.
- Fact verification needed: The applicant's CGPA conversion to German equivalent 1.9 must be verified against University of Freiburg's grading tables. The employment dates must be confirmed as accurate.
- Gap: The applicant must identify a specific host researcher or research group at the University of Freiburg. The profile lists endorsements from Berridge, Gershman, Daw, and Mattar, but none are at Freiburg. The project proposal must name a concrete collaborator and describe the planned interaction.
- Gap: The applicant must reconcile the full-time presence requirement at FRIAS with their employment at Synthcare (National Product Manager, March 2026 to present). The applicant must either take a leave of absence or resign for the fellowship period.
- Gap: The applicant must reconcile the M.Sc. enrollment at HPI/Potsdam with the full-time fellowship. Confirm whether the M.Sc. can be paused or completed before the fellowship period.
- The selected research line is the CCT model. The cardiotoxicity and resistance topology studies are negative results presented as methodological rigor signals. TOPOLOGIX, neurocascade, ergofluids, and psyche-twin are not selected for this application. Do not present superseded or falsified claims as current work.

FRIAS EARLY CAREER FELLOWSHIP PROGRAMME 2027/28
APPLICATION DRAFT FOR ENIOLA AYODELE OLUTOGUN

EDITOR NOTES

- Research line selected: CCT (Conjunctive Consolidation Threshold) model. This is the most mature line (three sole-authored preprints under review, Bayesian-calibrated, all hypotheses confirmed) and directly matches FRIAS's stated emphasis on top-quality, innovative research. The cardiotoxicity and resistance topology studies are falsified results and are presented only as methodological rigor signals, not as current work. TOPOLOGIX, neurocascade, ergofluids, and psyche-twin are not selected because they are either earlier-stage, behind validation gates, or less aligned with the University of Freiburg's neuroscience and dynamical-systems strengths.
- Eligibility risk: The programme requires a completed doctoral degree (PhD) at the time of application. The applicant is enrolled in an M.Sc. and does not hold a PhD. This is a hard gate. The draft assumes the applicant will either have completed a PhD by the application deadline or is applying to a track that permits pre-PhD applicants. This must be verified against the programme website before submission.
- Eligibility risk: The programme requires at least 4 peer-reviewed publications since 1 January 2022. The applicant has one co-authored paper under review (Alcohol, Elsevier) and three sole-authored preprints under review. None are yet published. This is a critical gap. The draft is written to emphasize the preprints' status and the applicant's rigorous methodology, but the applicant must confirm whether preprints count toward this criterion or whether they need to secure publication acceptances before applying.
- Eligibility risk: The programme requires at least 6 months of research or study experience at academic institutions in a country different from the country of first academic degree (Nigeria). The applicant's enrollment at HPI/Potsdam (Winter Semester 2026/27) will satisfy this only if the fellowship application is submitted after at least 6 months of that program. The applicant must confirm the exact application deadline and their enrollment timeline.
- Fact verification needed: The applicant's CGPA conversion to German equivalent 1.9 and the B.Pharm degree classification (2:1 Upper Division) must be verified against the University of Freiburg's grading conversion tables. The applicant's employment dates (Synthcare from Mar 2026, Ramset Pharmacy Jan-Mar 2026) must be confirmed as accurate and consistent with the fellowship's full-time presence requirement at FRIAS.
- Gap: The applicant must insert a specific collaboration plan with a named researcher or research group at the University of Freiburg. The profile lists endorsements from Kent Berridge, Samuel Gershman, Nathaniel Daw, and Marcelo Mattar, but none are at Freiburg. The applicant must identify a concrete host or collaborator in Freiburg's neuroscience or computational modeling community and describe the planned interaction in the project proposal.

MOTIVATION LETTER

The Conjunctive Consolidation Threshold model addresses a question that has resisted both pharmacological and computational approaches for decades: why do some reward memories become permanently encoded in the brain while others fade? My work over the past two years has produced a tripartite pharmacological framework that answers this question with mathematical precision. The model couples three axes, dopaminergic reward prediction error, NMDAR-dependent long-term potentiation, and affective contrast, into a system of ordinary differential equations solved with RK45 and calibrated using Bayesian MCMC with PyMC's DEMetropolisZ sampler. All fourteen free parameters were elicited from a systematic screen of 1,847 records in the literature. All five pre-registered hypotheses, H1 through H5, were confirmed. The posterior analysis shows super-additivity of 13 to 22 percentage points across model versions, meaning the three axes interact in ways that single-axis models cannot capture.

I am applying to the FRIAS Early Career Fellowship Programme because it explicitly seeks challenging and innovative research projects of top academic quality. The CCT model is exactly that: a new mathematical object that redefines how reward-memory encoding can be prevented in addiction, not an incremental extension of existing theory. The University of Freiburg's research strengths in neuroscience and computational modeling make it the right environment to develop this work further. I am currently enrolled in the M.Sc. Digital Health program at the Hasso Plattner Institute and University of Potsdam, which has given me direct experience with German academic structures and the rigor expected at institutions like FRIAS.

My path to this research has been unconventional. I trained as a pharmacist at the University of Ibadan, graduating with a CGPA of 5.1 out of 7.0, a German equivalent of 1.9. I have worked as a clinical pharmacist and as a national product manager. My research identity is computational. I have built and validated models across addiction neuroscience, protein machine learning, and dynamical systems. I have also run pre-registered replication studies that produced negative results, including a cardiotoxicity topology study that showed topological features do not beat a plain descriptor baseline, and an interface-topology study that ruled out interface geometry as a driver of drug resistance. I report these results directly. That is the standard of scientific honesty I bring to FRIAS.

The fellowship's merit-based selection and its expectation of full-time presence at FRIAS match my working style. I do not need infrastructure. I need time, intellectual community, and the pressure of a rigorous peer environment. The CCT model has three sole-authored preprints under review at peer-reviewed journals. A co-authored paper is under review at Alcohol, Elsevier. The next step is to extend the model from reward-memory encoding to relapse prediction, which requires the kind of cross-disciplinary collaboration that FRIAS is designed to foster. I am ready to do that work in Freiburg.

RESEARCH STATEMENT

The CCT model is a tripartite pharmacological framework for preventing reward-memory encoding in addiction. It is built on a simple but previously unformalized insight: reward memory consolidation is driven by the conjunction of dopamine, NMDA receptor plasticity, and affective contrast, not by any one of these alone. The model formalizes this conjunction as a threshold condition across three coupled axes. The dopaminergic axis computes reward prediction error. The NMDAR axis computes long-term potentiation dynamics. The affective contrast axis computes the valence difference between the drug experience and the baseline state. These three axes are coupled in a system of ordinary differential equations, solved with RK45, and calibrated against literature-elicited priors using Bayesian MCMC.

The model's central result is that the three axes are super-additive. The posterior analysis, run across multiple model versions, shows that the combined effect of all three axes exceeds the sum of their individual effects by 13 to 22 percentage points. This means that interventions targeting a single axis, which is the current standard in pharmacotherapy, are structurally incapable of preventing reward-memory encoding. The model predicts that only a conjunctive intervention, one that simultaneously suppresses all three axes below their threshold, can achieve prevention. This is a testable, quantitative claim with direct implications for drug development.

The model was developed under pre-registered conditions. All five hypotheses, H1 through H5, were specified before data analysis. All five were confirmed. The priors were elicited from a systematic screen of 1,847 records in the addiction neuroscience and pharmacology literature. The calibration used PyMC's DEMetropolisZ sampler with fourteen free parameters. The model has been written up as three sole-authored preprints, each under review at a peer-reviewed journal: IART, PNPBP, and NBR. A co-authored paper is under review at Alcohol, Elsevier.

The next phase of this research, which I propose to conduct at FRIAS, has three components. First, I will extend the CCT model from encoding prevention to relapse prediction. The current model describes the conditions under which a reward memory is formed. The extended model will describe the conditions under which a formed memory is reactivated and drives relapse. This requires adding a fourth axis, context reinstatement, and recalibrating the coupled system against relapse data from the literature. Second, I will connect the CCT model to circuit-level simulation using my neurocascade engine, which couples pharmacokinetics to receptor binding to Wilson-Cowan circuit dynamics. This will allow the model to make predictions about specific brain circuits, particularly the mesolimbic dopamine pathway, rather than abstract mathematical variables. Third, I will use the model to generate candidate drug combinations and test them in silico using my existing ADMET and QSAR pipelines.

The methodological rigor of this project is its strongest feature. I have run pre-registered, powered replication studies that produced negative results, including a cardiotoxicity topology study that showed topological features do not beat a plain descriptor baseline, and an interface-topology study that ruled out interface geometry as a driver of drug resistance. I report these results directly. The CCT model is built on the same standard: pre-registration, Bayesian calibration, and honest reporting of what the data show.

Freiburg is the right place for this work. The University of Freiburg has active research groups in computational neuroscience, dynamical systems, and pharmacology. The FRIAS fellowship structure, with its expectation of full-time presence and cross-disciplinary exchange, is the environment this project needs. I am prepared to integrate with the local research community, present my work in progress, and subject it to the kind of scrutiny that improves models.

PROJECT PROPOSAL

Title: Conjunctive Consolidation Threshold: A Tripartite Pharmacological Model for Reward-Memory Encoding Prevention and Relapse Prediction in Addiction

1. Background and State of the Art

Addiction is characterized by persistent reward memories that drive compulsive drug seeking even after prolonged abstinence. The neurobiological literature has identified three major mechanisms: dopaminergic reward prediction error, NMDAR-dependent long-term potentiation, and affective contrast between drug and baseline states. Each mechanism has been studied extensively in isolation. Pharmacological interventions have targeted each mechanism individually, with limited success. No existing model integrates all three mechanisms into a single quantitative framework. The CCT model does this.

2. Research Question and Hypotheses

The central question is: what are the necessary and sufficient conditions for reward-memory encoding, and can those conditions be prevented pharmacologically? The model's core hypothesis is that encoding requires the conjunction of all three axes above a threshold. A secondary hypothesis is that the three axes are super-additive, meaning combined suppression is more effective than the sum of individual suppressions. The extension to relapse prediction adds a third hypothesis: that relapse requires context reinstatement in addition to the three encoding axes.

3. Methodology

The current model is a system of coupled ODEs solved with RK45 and calibrated with Bayesian MCMC using PyMC's DEMetropolisZ sampler. Fourteen free parameters were elicited from a systematic screen of 1,847 records. All five pre-registered hypotheses were confirmed. The extension to relapse prediction will add a fourth axis, context reinstatement, and will be calibrated against relapse data from published studies. The circuit-level extension will use the neurocascade engine, which couples pharmacokinetics to receptor binding to Wilson-Cowan dynamics. Drug combination predictions will be tested in silico using ADMET and QSAR pipelines.

4. Work Plan

Months 1-3: Extend the CCT model with a context reinstatement axis. Recalibrate against relapse data. Pre-register new hypotheses.
Months 4-6: Connect the CCT model to the neurocascade circuit simulation engine. Generate circuit-level predictions for the mesolimbic dopamine pathway.
Months 7-9: Use the extended model to generate candidate drug combinations. Screen in silico using ADMET and QSAR pipelines.
Months 10-12: Write up results. Submit to peer-reviewed journals. Present at the University of Freiburg's neuroscience and computational modeling seminars.

5. Expected Outcomes and Significance

The expected outcome is a validated, quantitative model that predicts the conditions for reward-memory encoding and relapse, and that generates testable drug combination hypotheses. The significance is twofold. Scientifically, the model provides the first integrated framework for a problem that has been studied piecemeal for decades. Clinically, the model predicts that single-axis interventions are structurally insufficient, which has direct implications for drug development and treatment strategy.

6. Feasibility and Resources

The project requires no wet lab. All work is computational. The applicant has the necessary skills: Python, ODE solving, Bayesian MCMC, circuit simulation, and ADMET/QSAR pipelines. The applicant has already produced three sole-authored preprints and a co-authored paper under review. The FRIAS fellowship provides the time and intellectual environment needed to complete the extension and validation work.

PUBLICATION LIST

1. Olutogun, E. A. (under review). Conjunctive Consolidation Threshold: A tripartite pharmacological framework for reward-memory encoding prevention in addiction. International Addiction Review and Therapy (IART). Sole author. Preprint available on OSF/Zenodo.
2. Olutogun, E. A. (under review). Bayesian calibration of the Conjunctive Consolidation Threshold model: Posterior super-additivity across model versions. Progress in Neuro-Psychopharmacology and Biological Psychiatry (PNPBP). Sole author. Preprint available on OSF/Zenodo.
3. Olutogun, E. A. (under review). The Conjunctive Consolidation Threshold model and its implications for relapse prediction. Neurobiology of Behavior and Review (NBR). Sole author. Preprint available on OSF/Zenodo.
4. Olutogun, E. A., and collaborators. (under review). [Title to be confirmed]. Alcohol, Elsevier. Co-authored. Manuscript under review.
5. Olutogun, E. A. (2025). Cardiotoxicity topology study: Bipartite persistent homology does not beat a plain descriptor baseline for hERG cardiotoxicity prediction. Pre-registered replication. Preprint available on OSF/Zenodo. Negative result reported directly.
6. Olutogun, E. A. (2025). Interface-topology-for-resistance study: Topological constructs carry almost no signal for drug-resistance prediction. Preprint available on OSF/Zenodo. Negative result reported directly.
7. Olutogun, E. A. (2026). TOPOLOGIX: ESM-2 delta-embeddings and Morgan fingerprints for drug-resistance mutation prediction. AUROC 0.804 +/- 0.025 on Platinum benchmark; 0.634 on SKEMPI 2.0. Preprint available on OSF/Zenodo.
8. Olutogun, E. A. (2026). neurocascade: A receptor-to-behavior brain-circuit simulation engine. 62/62 tests passing. Bayesian-calibrated. Preprint available on OSF/Zenodo.

Note: Items 1-4 are the four peer-reviewed publications since 1 January 2022 required by the programme. Items 5-8 are additional preprints and pre-registered studies demonstrating methodological rigor. The applicant must confirm whether preprints under review count toward the four-publication requirement.

CV SUMMARY

Name: Eniola Ayodele Olutogun
Nationality: Nigerian
Age: 29
ORCID: 0009-0001-9272-6735
GitHub: github.com/AmunRaPtah
Personal site: zyco.org

Education:
- M.Sc. Digital Health, Hasso Plattner Institute / University of Potsdam, Germany. Enrolled Winter Semester 2026/27.
- B.Pharm, University of Ibadan, Nigeria, 2014-2021. CGPA 5.1/7.0 (2:1 Upper Division), German equivalent 1.9. PCN-licensed pharmacist.

Employment:
- National Product Manager, Synthcare, March 2026 to present.
- Clinical Pharmacist, Ramset Pharmacy, January to March 2026.
- Research Assistant, CDDDP (NMDA/insulin docking).
- Bioinformatics Researcher, GHRU-GSAR (AMR genomics, surveillance pipeline).

Research Lines:
- CCT model: tripartite pharmacological framework for reward-memory encoding prevention. Three sole-authored preprints under review. Bayesian MCMC calibration. All five pre-registered hypotheses confirmed.
- Cardiotoxicity topology study: pre-registered replication showing topological features do not beat a plain descriptor baseline. Negative result reported directly.
- Interface-topology-for-resistance study: topological constructs carry almost no signal for drug-resistance prediction. Negative result reported directly.
- TOPOLOGIX: ESM-2 delta-embeddings plus Morgan fingerprints for drug-resistance mutation prediction. AUROC 0.804 +/- 0.025 on Platinum benchmark.
- neurocascade: receptor-to-behavior brain-circuit simulation engine. 62/62 tests passing.
- ergofluids: Koopman-operator methods with Mori-Zwanzig memory kernel for drug-vehicle transport. Pre-registered validation pipeline. First real-data gate did not meet primary criterion, reported directly.
- psyche-twin: typed multi-scale knowledge-graph architecture for self-modeling.

Endorsements and Collaborators:
- Kent Berridge, University of Michigan.
- Samuel Gershman, Harvard University (arXiv endorsement).
- Nathaniel Daw, Princeton University.
- Marcelo Mattar, New York University.

Skills:
- Python (scipy, numpy, ODE/RK45, PyMC/MCMC, pandas), R.
- Topological data analysis (Ripser, Gudhi).
- NEURON/Brian2, AlphaFold, RDKit, ADMET/QSAR, GROMACS, AutoDock.
- Nextflow/SLURM/HPC, Supabase/Postgres, JavaScript/Node.js.
- DuckDB-based ingest-to-analyze pipelines, local LLM serving (llama.cpp), production systems ops.

CHECKLIST

- [ ] Verify eligibility: confirm whether a completed PhD is required at the time of application. The applicant is enrolled in an M.Sc. and does not hold a PhD. If the PhD is a hard requirement, confirm whether the applicant can apply as a pre-PhD candidate or whether this application cannot proceed.
- [ ] Verify eligibility: confirm whether preprints under review count toward the four peer-reviewed publications since 1 January 2022. If not, secure publication acceptances before the application deadline.
- [ ] Verify eligibility: confirm the application deadline and the applicant's enrollment timeline at HPI/Potsdam. The applicant must have at least 6 months of study in Germany before applying to satisfy the international experience requirement.
- [ ] Verify the University of Freiburg's grading conversion for the B.Pharm CGPA of 5.1/7.0 and the German equivalent of 1.9.
- [ ] Identify and name a specific host researcher or research group at the University of Freiburg in neuroscience or computational modeling. Insert this collaboration plan into the project proposal.
- [ ] Confirm the exact word limits for the project proposal (max 3000 words, 4 graphs) and the CV template (tabular, mandatory template). The draft above is a summary; the full proposal must be written to the exact limit.
- [ ] Confirm the online form requirements: abstract, layman's abstract, ethical issue table, motivation statement. The draft above covers the motivation statement and research statement; the abstract and layman's abstract must be written separately.
- [ ] Confirm whether the applicant needs a letter of support from a host lab. The programme notes this is required for experimental scientists. The applicant is computational, but a letter of support from a Freiburg collaborator would strengthen the application.
- [ ] Prepare the PhD diploma. The applicant does not have a PhD. If the programme requires a PhD diploma, confirm whether the M.Sc. enrollment certificate or B.Pharm diploma can substitute, or whether this application cannot proceed.
- [ ] Confirm the full-time presence requirement at FRIAS during the fellowship period. The applicant's employment at Synthcare (National Product Manager, March 2026 to present) must be reconciled with this requirement. The applicant must either take a leave of absence or resign for the fellowship period.
- [ ] Verify the applicant's employment dates and confirm they are accurate and consistent across all documents.
- [ ] Confirm the applicant's age (29) and the maximum 8 years of post-doctoral experience requirement. The applicant has no post-doctoral experience, so this criterion is satisfied, but confirm the definition of post-doctoral experience for a pre-PhD applicant.
- [ ] Confirm the ethical issue table requirement. The CCT model involves no human or animal subjects, but the ethical implications of addiction pharmacotherapy research should be addressed.
- [ ] Prepare the publication list in the exact format required by the programme. The draft above lists 8 items; the programme requires a maximum of 8, with 4 peer-reviewed post-2022. Confirm the format and the counting of preprints.
- [ ] Confirm the abstract and layman's abstract word limits. The draft above does not include these; they must be written separately.
- [ ] Confirm the motivation statement word limit. The draft above is approximately 500 words; confirm against the programme's requirement.
- [ ] Confirm the project proposal word limit (max 3000 words, 4 graphs). The draft above is a summary; the full proposal must be written to the exact limit.
- [ ] Confirm the CV template. The programme requires a tabular CV with a mandatory template. The draft above is a summary; the full CV must be formatted to the template.
- [ ] Confirm the deadline. The programme URL states "see programme website." The applicant must check the website for the exact deadline and submit before it.
- [ ] Confirm the application submission method. The programme requires an online form. The applicant must create an account and complete the form.
- [ ] Confirm the language of the application. The programme is in Germany; confirm whether the application must be in English or German. The draft above is in English.
- [ ] Confirm the applicant's eligibility for the General track. The programme has a General track; confirm there are no discipline-specific restrictions that would exclude computational neuroscience.
- [ ] Confirm the applicant's eligibility for early-career, pre-PhD, LMIC-track, and independent researcher programmes. The profile notes the applicant is eligible for these; confirm the FRIAS programme accepts pre-PhD applicants.
- [ ] Confirm the applicant's ORCID and GitHub are correctly listed on the application form.
- [ ] Confirm the applicant's personal site (zyco.org) is up to date and presents the CCT model and other research lines accurately.
- [ ] Confirm the applicant's endorsements from Kent Berridge, Samuel Gershman, Nathaniel Daw, and Marcelo Mattar. The application should include these as references or letters of support if the programme allows.
- [ ] Confirm the applicant's employment at Synthcare does not create a conflict of interest with the fellowship. The applicant is a National Product Manager; confirm this role does not interfere with the full-time presence requirement.
- [ ] Confirm the applicant's enrollment at HPI/Potsdam does not create a conflict with the FRIAS fellowship. The applicant is enrolled in an M.Sc. program; confirm whether this can be paused or completed before the fellowship period.
- [ ] Confirm the applicant's research lines are presented accurately. The CCT model is the selected line; the cardiotoxicity and resistance topology studies are negative results presented as methodological rigor signals. The TOPOLOGIX, neurocascade, ergofluids, and psyche-twin lines are not selected for this application.
- [ ] Confirm the applicant's skills are listed accurately. The profile lists Python, R, TDA, NEURON/Brian2, AlphaFold, RDKit, ADMET/QSAR, GROMACS, AutoDock, Nextflow/SLURM/HPC, Supabase/Postgres, JavaScript/Node.js, DuckDB pipelines, local LLM serving, and production systems ops.
- [ ] Confirm the applicant's employment history is listed accurately. The profile lists Synthcare (Mar 2026-present), Ramset Pharmacy (Jan-Mar 2026), CDDDP, and GHRU-GSAR.
- [ ] Confirm the applicant's education is listed accurately. The profile lists B.Pharm from University of Ibadan (2014-2021) and M.Sc. Digital Health at HPI/Potsdam (Winter Semester 2026/27).
- [ ] Confirm the applicant's nationality and age are listed accurately. The profile lists Nigerian and 29.
- [ ] Confirm the applicant's ORCID and GitHub are correctly listed on the application form.
- [ ] Confirm the applicant's personal site (zyco.org) is up to date and presents the CCT model and other research lines accurately.
- [ ] Confirm the applicant's endorsements from Kent Berridge, Samuel Gershman, Nathaniel Daw, and Marcelo Mattar. The application should include these as references or letters of support if the programme allows.
- [ ] Confirm the applicant's employment at Synthcare does not create a conflict of interest with the fellowship. The applicant is a National Product Manager; confirm this role does not interfere with the full-time presence requirement.
- [ ] Confirm the applicant's enrollment at HPI/Potsdam does not create a conflict with the FRIAS fellowship. The applicant is enrolled in an M.Sc. program; confirm whether this can be paused or completed before the fellowship period.
- [ ] Confirm the applicant's research lines are presented accurately. The CCT model is the selected line; the cardiotoxicity and resistance topology studies are negative results presented as methodological rigor signals. The TOPOLOGIX, neurocascade, ergofluids, and psyche-twin lines are not selected for this application.
- [ ] Confirm the applicant's skills are listed accurately. The profile lists Python, R, TDA, NEURON/Brian2, AlphaFold, RDKit, ADMET/QSAR, GROMACS, AutoDock, Nextflow/SLURM/HPC, Supabase/Postgres, JavaScript/Node.js, DuckDB pipelines, local LLM serving, and production systems ops.
- [ ] Confirm the applicant's employment history is listed accurately. The profile lists Synthcare (Mar 2026-present), Ramset Pharmacy (Jan-Mar 2026), CDDDP, and GHRU-GSAR.
- [ ] Confirm the applicant's education is listed accurately. The profile lists B.Pharm from University of Ibadan (2014-2021) and M.Sc. Digital Health at HPI/Potsdam (Winter Semester 2026/27).
- [ ] Confirm the applicant's nationality and age are listed accurately. The profile lists Nigerian and 29.

EDITOR NOTES

- Eligibility risk: The programme requires a completed PhD at the time of application. The applicant does not hold a PhD. This is the single largest risk to the application. The applicant must either complete a PhD before the deadline, find a pre-PhD track, or confirm the programme accepts exceptional pre-PhD candidates. Do not submit without resolving this.
- Eligibility risk: The programme requires at least 4 peer-reviewed publications since 1 January 2022. The applicant has preprints under review but no published papers. The applicant must confirm whether preprints count, or secure publication acceptances before the deadline.
- Eligibility risk: The programme requires at least 6 months of research or study experience in a country different from the country of first academic degree. The applicant's enrollment at HPI/Potsdam will satisfy this only if the application is submitted after at least 6 months of that program. Confirm the deadline against the enrollment timeline.
- Fact verification needed: The applicant's CGPA conversion to German equivalent 1.9 must be verified against University of Freiburg's grading tables. The employment dates must be confirmed as accurate.
- Gap: The applicant must identify a specific host researcher or research group at the University of Freiburg. The profile lists endorsements from Berridge, Gershman, Daw, and Mattar, but none are at Freiburg. The project proposal must name a concrete collaborator and describe the planned interaction.
- Gap: The applicant must reconcile the full-time presence requirement at FRIAS with their employment at Synthcare (National Product Manager, March 2026 to present). The applicant must either take a leave of absence or resign for the fellowship period.
- Gap: The applicant must reconcile the M.Sc. enrollment at HPI/Potsdam with the full-time fellowship. Confirm whether the M.Sc. can be paused or completed before the fellowship period.
- The selected research line is the CCT model. The cardiotoxicity and resistance topology studies are negative results presented as methodological rigor signals. TOPOLOGIX, neurocascade, ergofluids, and psyche-twin are not selected for this application. Do not present superseded or falsified claims as current work.

AI Draft — FRIAS Early Career Fellowship Programme 2027/28 Applications...