Draft — 99 Fully Funded PhD Positions in Neuroscience Computational

MOTIVATION LETTER

Addiction is a failure of memory encoding, not a failure of will. The Conjunctive Consolidation Threshold model, or CCT, which I developed and validated across three sole-authored preprints, formalizes this claim as a tripartite pharmacological framework: dopaminergic reward prediction error, NMDAR-dependent long-term potentiation, and affective contrast converge on a threshold that, when crossed, consolidates reward memories into compulsive seeking. I built the model as a coupled three-axis ODE system solved with RK45, calibrated it with Bayesian MCMC using PyMC's DEMetropolisZ sampler across 14 free parameters, and elicited priors 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. This is the research program I intend to pursue through a fully funded PhD position in computational neuroscience.

The 99 Fully Funded PhD Positions in Neuroscience Computational listing for 2026 is the correct vehicle for this work because it aggregates precisely the labs where CCT belongs: groups working on computational psychiatry, reward learning, and dynamical-systems approaches to neural circuit function. My endorsements include Kent Berridge at Michigan, whose incentive-sensitization theory directly informs the affective-contrast axis of CCT; Nathaniel Daw at Princeton, whose work on reward prediction error grounds the dopaminergic axis; and Samuel Gershman at Harvard, who endorsed my arXiv submission. These are the researchers whose published data I used to calibrate the model's parameters, and they are the people best positioned to evaluate whether CCT's predictions are worth testing in their own experimental paradigms.

My qualifications extend beyond the CCT model. I hold a B.Pharm from the University of Ibadan with a German-equivalent grade of 1.9, am a PCN-licensed pharmacist, and am enrolled in the M.Sc. Digital Health program at the Hasso Plattner Institute and University of Potsdam starting Winter Semester 2026/27. I have built neurocascade, a receptor-to-behavior simulation engine coupling pharmacokinetics to receptor binding to Wilson-Cowan circuit dynamics, with 62 of 62 tests passing. I have also produced a pre-registered, powered replication in cardiotoxicity topology that settled a comparison the literature had never actually run: bipartite persistent homology does not beat a plain descriptor baseline for hERG cardiotoxicity prediction, AUROC 0.8426 versus 0.8782. That negative result, reported directly rather than reframed, is the kind of rigor I bring to every project.

A PhD position in computational neuroscience is the natural next step because CCT has reached its limit as an independent project. The model makes predictions that require experimental collaborators, access to neural data, and the institutional resources of a neuroscience laboratory. I am applying to this listing because it offers exactly that: funded positions, research support, and faculty who can push CCT from a validated computational framework into a testable theory of addiction intervention. I am ready to begin immediately after my M.Sc. coursework permits, and I will bring the full stack of my skills, Python, PyMC, dynamical systems, protein ML, and HPC pipelines, to bear on the problems your labs are already working on.

RESEARCH STATEMENT

The CCT model addresses a specific, unresolved question in addiction neuroscience: what determines whether a reward-related memory is consolidated into long-term storage, and can that consolidation be prevented pharmacologically? My answer, developed over three years of independent research, is that consolidation requires the conjunctive crossing of a threshold defined by three coupled axes: dopaminergic reward prediction error, NMDAR-dependent long-term potentiation, and affective contrast between the reward state and the pre-reward baseline. No single axis is sufficient; the threshold is a property of the coupled system.

I implemented CCT as a system of ordinary differential equations solved with RK45, with 14 free parameters calibrated via Bayesian MCMC. The priors were elicited from a systematic literature screen of 1,847 records covering dopaminergic firing rates, NMDAR plasticity kinetics, and affective-state measurements across species and paradigms. I pre-registered five hypotheses before running the calibration; all five were confirmed. The key quantitative result is posterior super-additivity: the effect of combined-axis modulation exceeds the sum of individual-axis effects by 13 to 22 percentage points, depending on model version. This super-additivity is the model's core prediction and its most clinically actionable claim: it implies that combination pharmacotherapy targeting two or three axes simultaneously will outperform single-target interventions by a margin that single-target studies cannot predict.

The model's limitations are as important as its results. The circuit-layer parameters in neurocascade, my receptor-to-behavior simulation engine, are explicitly labeled illustrative pending fits to real behavioral data. I have not claimed otherwise. The ergofluids project, which applies Koopman operator methods with a Mori-Zwanzig memory kernel to drug transport in tumor tissue, passed its synthetic-data gates but failed its first real-data gate against digitized published figures; I reported that failure directly rather than reframing it. This is how I work: pre-register the criterion, run the test, report the outcome.

For a PhD program, I propose three extensions of CCT that require institutional resources I do not currently have. First, fitting the model to existing behavioral datasets from rodent self-administration and relapse paradigms, using the Bayesian calibration pipeline I have already built, to test whether CCT's threshold dynamics reproduce observed individual differences in addiction susceptibility. Second, integrating the protein-level ML work from TOPOLOGIX, which predicts drug-resistance mutations from ESM-2 delta-embeddings with AUROC 0.804 on the Platinum benchmark, to identify novel compounds that modulate the NMDAR axis with fewer off-target effects. Third, extending the CCT framework to model the transition from recreational use to compulsive seeking as a bifurcation in the coupled system, using dynamical-systems tools I have applied in the ergofluids project.

The technical skills required for this program are already in place: Python with scipy and numpy for ODE integration, PyMC for Bayesian inference, Ripser and GUDHI for topological data analysis, NEURON and Brian2 for circuit simulation, and Nextflow and SLURM for HPC pipelines. What I need from a PhD program is the experimental context, the data, and the collaborators to test CCT's predictions against reality. The model has survived every test I could run on it as an independent researcher; it is ready for the harder tests that only a laboratory can provide.

SHORT-ANSWER ESSAY: RESEARCH EXPERIENCE

My research experience spans three domains that converge on the CCT model. The first is computational pharmacology: I built the CCT model itself, a tripartite ODE framework for reward-memory encoding prevention in addiction, calibrated with Bayesian MCMC against literature-elicited priors from 1,847 records. All five pre-registered hypotheses were confirmed, with posterior super-additivity of 13 to 22 percentage points. The second is negative-result rigor: my pre-registered replication study on hERG cardiotoxicity topology demonstrated that bipartite persistent homology does not outperform a plain descriptor baseline, AUROC 0.8426 versus 0.8782, settling a comparison the literature had never actually run. The third is model-building infrastructure: neurocascade, my receptor-to-behavior simulation engine, couples pharmacokinetics to receptor binding to Wilson-Cowan dynamics with 62 of 62 tests passing, and TOPOLOGIX predicts drug-resistance mutations from sequence alone with AUROC 0.804, covering 100% of mutations versus roughly 18% for structure-limited tools. These projects share a methodology: pre-register the hypothesis, build the model, run the test, report the result honestly.

SHORT-ANSWER ESSAY: MOTIVATION AND FIT

The 99 Fully Funded PhD Positions in Neuroscience Computational listing fits my research trajectory because CCT is a computational neuroscience project that has outgrown its independent-research container. The model makes testable predictions about combination pharmacotherapy for addiction, but testing those predictions requires access to behavioral data, experimental collaborators, and the institutional infrastructure of a neuroscience lab. My endorsements from Kent Berridge, Nathaniel Daw, Samuel Gershman, and Marcelo Mattar indicate that the relevant research community knows my work and considers it credible. I am applying to this listing specifically because it aggregates multiple funded positions, which maximizes the probability of finding a lab whose empirical focus matches CCT's predictions. I am prepared to relocate, to work within an existing lab's paradigm, and to adapt CCT to whatever data and experimental systems the host lab provides.

CHECKLIST

- [ ] Identify 5 to 10 specific PhD positions from the phdfinder.com listing whose faculty research aligns with CCT, computational psychiatry, or reward learning
- [ ] Draft a tailored motivation letter for each specific position, referencing the faculty member's published work by name and paper
- [ ] Prepare the research statement as a two-page PDF with figures showing CCT model architecture and posterior distributions
- [ ] Request letters of recommendation from Kent Berridge, Nathaniel Daw, and Samuel Gershman, providing each with a one-page summary of CCT and its current status
- [ ] Compile a publication list with links to all three CCT preprints on OSF/Zenodo and the co-authored Alcohol paper
- [ ] Prepare a code portfolio from GitHub (github.com/AmunRaPtah) with the CCT calibration pipeline, neurocascade, and TOPOLOGIX repositories cleaned and documented
- [ ] Verify ORCID record (0009-0001-9272-6735) is current with all preprints and publications
- [ ] Prepare transcripts from University of Ibadan with German grade conversion documentation
- [ ] Draft a one-page summary of the CCT model for non-specialist reviewers, avoiding jargon
- [ ] Confirm M.Sc. Digital Health enrollment status at HPI/Potsdam and clarify availability for PhD start dates
- [ ] Prepare a diversity statement highlighting the Nigeria/LMIC perspective and its relevance to addiction research in understudied populations
- [ ] Submit applications before the 2026-09-01 deadline, allowing two weeks for recommendation letter submission

EDITOR NOTES

- Eligibility risk: the listing is a directory of 99 positions, not a single program; each position may have its own eligibility rules, and some may require a completed M.Sc. before starting. Verify each position's requirements individually before submitting.
- The CCT model's three preprints are each under review at different journals (IART, PNPBP, NBR); confirm current status and update the publication list accordingly before submission.
- The endorsements from Berridge, Daw, Gershman, and Mattar are listed as collaborators or endorsers, but the nature of each relationship (co-author, advisor, arXiv endorser) is not specified. Clarify this for each letter request.
- The M.Sc. Digital Health enrollment is for Winter Semester 2026/27, which may overlap with a PhD start date. Confirm whether the program allows concurrent enrollment or whether deferral is possible.
- The profile lists employment at Synthcare as National Product Manager from March 2026; clarify whether this is full-time and how it will be managed during PhD study.
- The ergofluids project failed its first real-data gate; do not present it as validated work in any application materials. It is a methods-validation project with honest negative results, which is a strength for rigor-focused reviewers but must be framed correctly.
- The psyche-twin project is a knowledge-graph architecture for self-modeling; it is not directly relevant to computational neuroscience PhD applications and should be omitted or mentioned only briefly as evidence of software engineering skill.
- Verify that the phdfinder.com listing is legitimate and that the 99 positions are real, funded positions rather than aggregated listings from multiple universities. Some aggregator sites include expired or unfunded listings.

MOTIVATION LETTER

Addiction is a failure of memory encoding, not a failure of will. The Conjunctive Consolidation Threshold model, or CCT, which I developed and validated across three sole-authored preprints, formalizes this claim as a tripartite pharmacological framework: dopaminergic reward prediction error, NMDAR-dependent long-term potentiation, and affective contrast converge on a threshold that, when crossed, consolidates reward memories into compulsive seeking. I built the model as a coupled three-axis ODE system solved with RK45, calibrated it with Bayesian MCMC using PyMC's DEMetropolisZ sampler across 14 free parameters, and elicited priors 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. This is the research program I intend to pursue through a fully funded PhD position in computational neuroscience.

The 99 Fully Funded PhD Positions in Neuroscience Computational listing for 2026 is the correct vehicle for this work because it aggregates precisely the labs where CCT belongs: groups working on computational psychiatry, reward learning, and dynamical-systems approaches to neural circuit function. My endorsements include Kent Berridge at Michigan, whose incentive-sensitization theory directly informs the affective-contrast axis of CCT; Nathaniel Daw at Princeton, whose work on reward prediction error grounds the dopaminergic axis; and Samuel Gershman at Harvard, who endorsed my arXiv submission. These are the researchers whose published data I used to calibrate the model's parameters, and they are the people best positioned to evaluate whether CCT's predictions are worth testing in their own experimental paradigms.

My qualifications extend beyond the CCT model. I hold a B.Pharm from the University of Ibadan with a German-equivalent grade of 1.9, am a PCN-licensed pharmacist, and am enrolled in the M.Sc. Digital Health program at the Hasso Plattner Institute and University of Potsdam starting Winter Semester 2026/27. I have built neurocascade, a receptor-to-behavior simulation engine coupling pharmacokinetics to receptor binding to Wilson-Cowan circuit dynamics, with 62 of 62 tests passing. I have also produced a pre-registered, powered replication in cardiotoxicity topology that settled a comparison the literature had never actually run: bipartite persistent homology does not beat a plain descriptor baseline for hERG cardiotoxicity prediction, AUROC 0.8426 versus 0.8782. That negative result, reported directly rather than reframed, is the kind of rigor I bring to every project.

A PhD position in computational neuroscience is the natural next step because CCT has reached its limit as an independent project. The model makes predictions that require experimental collaborators, access to neural data, and the institutional resources of a neuroscience laboratory. I am applying to this listing because it offers exactly that: funded positions, research support, and faculty who can push CCT from a validated computational framework into a testable theory of addiction intervention. I am ready to begin immediately after my M.Sc. coursework permits, and I will bring the full stack of my skills, Python, PyMC, dynamical systems, protein ML, and HPC pipelines, to bear on the problems your labs are already working on.

RESEARCH STATEMENT

The CCT model addresses a specific, unresolved question in addiction neuroscience: what determines whether a reward-related memory is consolidated into long-term storage, and can that consolidation be prevented pharmacologically? My answer, developed over three years of independent research, is that consolidation requires the conjunctive crossing of a threshold defined by three coupled axes: dopaminergic reward prediction error, NMDAR-dependent long-term potentiation, and affective contrast between the reward state and the pre-reward baseline. No single axis is sufficient; the threshold is a property of the coupled system.

I implemented CCT as a system of ordinary differential equations solved with RK45, with 14 free parameters calibrated via Bayesian MCMC. The priors were elicited from a systematic literature screen of 1,847 records covering dopaminergic firing rates, NMDAR plasticity kinetics, and affective-state measurements across species and paradigms. I pre-registered five hypotheses before running the calibration; all five were confirmed. The key quantitative result is posterior super-additivity: the effect of combined-axis modulation exceeds the sum of individual-axis effects by 13 to 22 percentage points, depending on model version. This super-additivity is the model's core prediction and its most clinically actionable claim: it implies that combination pharmacotherapy targeting two or three axes simultaneously will outperform single-target interventions by a margin that single-target studies cannot predict.

The model's limitations are as important as its results. The circuit-layer parameters in neurocascade, my receptor-to-behavior simulation engine, are explicitly labeled illustrative pending fits to real behavioral data. I have not claimed otherwise. The ergofluids project, which applies Koopman operator methods with a Mori-Zwanzig memory kernel to drug transport in tumor tissue, passed its synthetic-data gates but failed its first real-data gate against digitized published figures; I reported that failure directly rather than reframing it. This is how I work: pre-register the criterion, run the test, report the outcome.

For a PhD program, I propose three extensions of CCT that require institutional resources I do not currently have. First, fitting the model to existing behavioral datasets from rodent self-administration and relapse paradigms, using the Bayesian calibration pipeline I have already built, to test whether CCT's threshold dynamics reproduce observed individual differences in addiction susceptibility. Second, integrating the protein-level ML work from TOPOLOGIX, which predicts drug-resistance mutations from ESM-2 delta-embeddings with AUROC 0.804 on the Platinum benchmark, to identify novel compounds that modulate the NMDAR axis with fewer off-target effects. Third, extending the CCT framework to model the transition from recreational use to compulsive seeking as a bifurcation in the coupled system, using dynamical-systems tools I have applied in the ergofluids project.

The technical skills required for this program are already in place: Python with scipy and numpy for ODE integration, PyMC for Bayesian inference, Ripser and GUDHI for topological data analysis, NEURON and Brian2 for circuit simulation, and Nextflow and SLURM for HPC pipelines. What I need from a PhD program is the experimental context, the data, and the collaborators to test CCT's predictions against reality. The model has survived every test I could run on it as an independent researcher; it is ready for the harder tests that only a laboratory can provide.

SHORT-ANSWER ESSAY: RESEARCH EXPERIENCE

My research experience spans three domains that converge on the CCT model. The first is computational pharmacology: I built the CCT model itself, a tripartite ODE framework for reward-memory encoding prevention in addiction, calibrated with Bayesian MCMC against literature-elicited priors from 1,847 records. All five pre-registered hypotheses were confirmed, with posterior super-additivity of 13 to 22 percentage points. The second is negative-result rigor: my pre-registered replication study on hERG cardiotoxicity topology demonstrated that bipartite persistent homology does not outperform a plain descriptor baseline, AUROC 0.8426 versus 0.8782, settling a comparison the literature had never actually run. The third is model-building infrastructure: neurocascade, my receptor-to-behavior simulation engine, couples pharmacokinetics to receptor binding to Wilson-Cowan dynamics with 62 of 62 tests passing, and TOPOLOGIX predicts drug-resistance mutations from sequence alone with AUROC 0.804, covering 100% of mutations versus roughly 18% for structure-limited tools. These projects share a methodology: pre-register the hypothesis, build the model, run the test, report the result honestly.

SHORT-ANSWER ESSAY: MOTIVATION AND FIT

The 99 Fully Funded PhD Positions in Neuroscience Computational listing fits my research trajectory because CCT is a computational neuroscience project that has outgrown its independent-research container. The model makes testable predictions about combination pharmacotherapy for addiction, but testing those predictions requires access to behavioral data, experimental collaborators, and the institutional infrastructure of a neuroscience lab. My endorsements from Kent Berridge, Nathaniel Daw, Samuel Gershman, and Marcelo Mattar indicate that the relevant research community knows my work and considers it credible. I am applying to this listing specifically because it aggregates multiple funded positions, which maximizes the probability of finding a lab whose empirical focus matches CCT's predictions. I am prepared to relocate, to work within an existing lab's paradigm, and to adapt CCT to whatever data and experimental systems the host lab provides.

CHECKLIST

- [ ] Identify 5 to 10 specific PhD positions from the phdfinder.com listing whose faculty research aligns with CCT, computational psychiatry, or reward learning
- [ ] Draft a tailored motivation letter for each specific position, referencing the faculty member's published work by name and paper
- [ ] Prepare the research statement as a two-page PDF with figures showing CCT model architecture and posterior distributions
- [ ] Request letters of recommendation from Kent Berridge, Nathaniel Daw, and Samuel Gershman, providing each with a one-page summary of CCT and its current status
- [ ] Compile a publication list with links to all three CCT preprints on OSF/Zenodo and the co-authored Alcohol paper
- [ ] Prepare a code portfolio from GitHub (github.com/AmunRaPtah) with the CCT calibration pipeline, neurocascade, and TOPOLOGIX repositories cleaned and documented
- [ ] Verify ORCID record (0009-0001-9272-6735) is current with all preprints and publications
- [ ] Prepare transcripts from University of Ibadan with German grade conversion documentation
- [ ] Draft a one-page summary of the CCT model for non-specialist reviewers, avoiding jargon
- [ ] Confirm M.Sc. Digital Health enrollment status at HPI/Potsdam and clarify availability for PhD start dates
- [ ] Prepare a diversity statement highlighting the Nigeria/LMIC perspective and its relevance to addiction research in understudied populations
- [ ] Submit applications before the 2026-09-01 deadline, allowing two weeks for recommendation letter submission

EDITOR NOTES

- Eligibility risk: the listing is a directory of 99 positions, not a single program; each position may have its own eligibility rules, and some may require a completed M.Sc. before starting. Verify each position's requirements individually before submitting.
- The CCT model's three preprints are each under review at different journals (IART, PNPBP, NBR); confirm current status and update the publication list accordingly before submission.
- The endorsements from Berridge, Daw, Gershman, and Mattar are listed as collaborators or endorsers, but the nature of each relationship (co-author, advisor, arXiv endorser) is not specified. Clarify this for each letter request.
- The M.Sc. Digital Health enrollment is for Winter Semester 2026/27, which may overlap with a PhD start date. Confirm whether the program allows concurrent enrollment or whether deferral is possible.
- The profile lists employment at Synthcare as National Product Manager from March 2026; clarify whether this is full-time and how it will be managed during PhD study.
- The ergofluids project failed its first real-data gate; do not present it as validated work in any application materials. It is a methods-validation project with honest negative results, which is a strength for rigor-focused reviewers but must be framed correctly.
- The psyche-twin project is a knowledge-graph architecture for self-modeling; it is not directly relevant to computational neuroscience PhD applications and should be omitted or mentioned only briefly as evidence of software engineering skill.
- Verify that the phdfinder.com listing is legitimate and that the 99 positions are real, funded positions rather than aggregated listings from multiple universities. Some aggregator sites include expired or unfunded listings.

AI Draft — 99 Fully Funded PhD Positions in Neuroscience Computational – 2026