PhD Project: Opening up the decision-making process in the mind and the brain

PhD Project: Opening up the decision-making process in the mind and the brain

This doctorate asks whether a sampling account of human judgments can be connected to the neural activity underlying decisions. It combines behavioural think-aloud work, neural decoding and computational modelling, rather than merely predicting a final choice. Adam Sanborn leads the Warwick project with Nick Chater and Aston's Alexander Zhigalov. The methodological opportunity is strong, but similar temporal statistics do not establish a unique neural algorithm. Warwick covers the overseas tuition difference for successful international MIBTP awardees; funding and the preferred project remain competitive. The MIBTP round for October 2027 entry lists a 23 November 2026 deadline; the opening date, closing time and time zone are not stated, and current acceptance status is unconfirmed. Confirm the exact course/English conditions and neural-data arrangements before investing in an application. Official project; current application and funding.

Research Problem and Doctoral Work

The advertised research starts from a distinction between the answer a person gives and the sequence of internal steps producing it. Face-age estimates and credit-card repayments illustrate uncertainty and anchoring; they are not a commitment that this student will run those exact tasks. Bayesian models specify how beliefs should relate to evidence, while a process model must explain how a limited cognitive system reaches a response. The team proposes that locally generated, autocorrelated samples can explain several otherwise puzzling biases and fluctuations.

The doctorate's advertised empirical tools are behavioural think-aloud procedures and neural decoding, combined with new data and computational modelling. The scientific target is the currently unknown connection between behavioural sampling algorithms and their neural substrate. The announcement does not assign a particular EEG/MEG scanner, brain region, participant count, clinical population, sampling rate or fixed experimental protocol. A collaborator's expertise or a previous paper's equipment cannot fill those gaps. This is basic cognitive/neuroscience research, with no advertised diagnostic or treatment outcome.

A useful interpretation is to examine how a time-dependent sequence constrains competing explanations. A model can reproduce average choices while missing autocorrelation, large jumps or response timing. Conversely, several algorithms or observation rules can generate similar statistics. The work therefore requires attention to what is sampled, how an internal sample becomes a verbal response or neural measurement, and whether the available observations distinguish the alternatives. These are analytical questions for project evaluation, not additional official admission tests or a prescribed study design.

Think-aloud reports also need an explicit interpretation: verbalisation may reveal only part of the process and may change task demands. Successful neural prediction would show information in a measured signal; it would not by itself identify a unique algorithm or establish causality. Ask how behavioural, speech and neural evidence will be coordinated, what alternative explanations will be compared, and which outcomes remain useful if the expected correspondence is absent. A defensible negative or non-identifying result can still contribute if the question is well bounded.

The 2022 cognitive-noise study and 2024 forecasting study show why variability is a research target rather than something to average away. Their model comparisons also expose dependence on representation, readout and the selected candidate set. The neural-source study offers a complementary caution: an apparently propagating sensor pattern need not uniquely reveal a cortical travelling process. The project's value lies in testing across these levels, rather than assuming their equivalence.

The Behavioural Science Group links mathematical modelling with experimental decision research, while MIBTP training offers quantitative development and non-academic exposure. Statistical modelling, human-data collection and reproducible analysis can transfer to cognitive research, behavioural analysis and quantitative research roles. These are skill-based interpretations, not verified project-specific job counts or salaries. A research PhD is not advertised here as clinical-practice training, and the repayment example does not establish financial-product or investment performance.

School, Programme and Field Context

University of Warwick is the explicit university of registration and intended doctoral awarding institution. Sanborn's academic home is the Department of Psychology; Chater contributes from Warwick Business School, and Zhigalov is currently a Lecturer in Biomedical Engineering at Aston University, confirmed by his institutional profile. Their collaboration does not establish an Aston degree, joint award or paid employment at either institution. The formal doctoral award title and the relationship between the research department and administrative course should be confirmed in the offer.

This distinction matters because the Warwick MIBTP instructions prescribe an application through School of Life Sciences, MIBTP DTP, course P-C1PBB. Applicants should follow that specific route rather than silently substituting a general Psychology PhD or WBS programme. The exact project appears in the current Neuroscience and Behaviour directory; a theme total is not an additional allocation of places to this topic.

Warwick is joint 68 in QS World University Rankings 2027. This overall ranking supports an institutional visibility assessment, not a Psychology subject ranking or supervision-quality measure; no verified ranking for this specific computational-cognition group is established here. The more directly relevant environment is the Behavioural Science Group: mathematical and behavioural research, weekly joint research meetings and DR@W seminars connecting Psychology, Economics, Mathematics and WBS. Its broad facilities account does not guarantee this student's access to a particular neural system or specialist technician.

The programme FAQ describes a four-year MIBTP framework across six partner universities. The detailed training outline remains headed 2025, despite its current availability, so its dates are a model rather than a confirmed 2027 timetable. It includes programming/statistics, later bioinformation and modelling, twelve masterclass days, annually reviewed development plans, and a three-month research-training placement at a non-home partner institution. A separate three-month non-academic PIPS placement is expected during the four years.

Technical visits need clarification: the FAQ describes a three-month mini-project, while the detailed outline allows a one-day to three-month visit in year two or three. Neither should be conflated with the first-year research placement or automatically added as a fixed third three-month absence. Quantitative training and cross-bioscience modules may be broader than this cognitive project. Confirm how they interact with participant recruitment, neural-data access and continuous modelling work.

The supervisor's weekly meetings and training expectations are useful, but are not a substitute for formal university progression, credits, upgrade, thesis, examination or teaching requirements. Those individual requirements and the 2027 handbook were not fully verified. The four-year training model also does not establish the exact months of an individual's funded award, permitted extensions or completion fees.

Supervisors, Lab and Recent Original Research

The official project identifies Professor Adam Sanborn as primary supervisor, and Professor Nick Chater and Dr Alexander Zhigalov as secondary supervisors. The project contact is a.n.sanborn@warwick.ac.uk. Sanborn's programme investigates whether MCMC approximations can provide a consistent account across cognitive tasks and response types. The sampling group account places him in Psychology and Chater in WBS, and supplies primary descriptions of their computational/empirical work. It is a research connection, not a guarantee that every linked tool will be maintained for the student.

Zhigalov’s Aston profile supplies the complementary neural expertise: EEG/MEG, signal processing, machine learning and long-term temporal associations. His earlier Birmingham employment explains links to that research environment, but does not establish the present project's measurement site or access rights. Ask who will train and supervise acquisition, decoding, behavioural experiments and model development, and how cross-institution visits and costs will work.

Sanborn's published supervision statement promises at least weekly scheduled meetings, using a mixture of face-to-face, video or telephone contact. Hours are flexible around agreed meetings; Slack communication normally occurs during the working week. Written work of up to 5,000 words needs at least one week's notice for feedback. The statement offers technical help from the supervisor and lab colleagues. These are declared expectations, not an independently measured completion rate, guaranteed instant availability or proof of supervision quality.

Zhu and colleagues (2022): cognitive noise

Zhu, León-Villagrá, Chater and Sanborn (2022), PLOS Computational Biology 18(8):e1010312; DOI 10.1371/journal.pcbi.1010312. The original cognitive-noise study combines animal naming and time estimation with approximate Bayesian model comparison. Heavy-tailed changes and long-range dependence co-occur, but algorithm ranking is task dependent: Random Walk Metropolis (RWM) and Metropolis-coupled MCMC (MC3) are comparable in animal naming, while MC3 better fits timing. The mapping from internal samples to observed responses changes the comparison. Ten naming participants and thirty analysed timing participants limit broad population extrapolation. This is behavioural evidence for computational hypotheses, not direct neural implementation. Reading: selected original HTML methods, results and discussion/conclusion; not all supporting information or raw data.

Spicer and colleagues (2024): forecasting

Spicer, Zhu, Chater and Sanborn (2024), Psychological Review 131(5):1069–1113; DOI 10.1037/rev0000493. The original forecasting paper compares model families against an existing untrended dataset and a new trended experiment, using approximate likelihood, model recovery and random-forest classification. Sampling models receive substantial relative support, but the winners and assignments vary with comparison method and statistics. Their psychophysical kernels and deviations remain imperfect, and the candidate set is not exhaustive. The new analysed sample contains 43 participants and was not preregistered. This teaches model discrimination, not that a classifier has identified an individual's brain algorithm. Reading: selected original PDF task/comparison methods, results and discussion; not every appendix or supplementary analysis.

Zhu and colleagues (2024): unified sampling

Zhu, Sundh, Spicer, Chater and Sanborn (2024), Psychological Review 131(2):456–493; DOI 10.1037/rev0000427. The ABS institutional record describes a single autocorrelated sampling process spanning six response families, including estimates, choice, confidence and response times. The publication is in the 2024 volume after 2023 online publication. Its contribution is a unifying qualitative framework; the hypothesis that cognitive variability comes from posterior sampling is not a proven neural fact. Reading: complete primary institutional abstract/metadata and indexed theoretical introduction only. Detailed model derivations and complete empirical comparisons were not examined here.

Zhigalov and Jensen (2023): neural sources

Zhigalov and Jensen (2023), NeuroImage 272:120047; DOI 10.1016/j.neuroimage.2023.120047. The primary institutional abstract reports MEG measurements, temporal-response analysis, source modelling and EEG/MEG forward models. Two phase-lagged discrete sources could explain apparent travelling echoes at scalp sensors, illustrating non-unique neural interpretation. It does not rule out intracortical travelling waves, and is not a sampling-algorithm test. Reading: complete original institutional abstract/metadata only; detailed controls were not examined. Equipment and computing acknowledged in that older study are not this student's resource entitlement.

These four recent original papers cover behavioural model comparison, theory and a neural interpretation problem. Two received selected main-text examination; the other two retain explicit abstract-level limits. They support an active, complementary research combination without resolving the advertised behavioural-to-neural link. The latter two papers cannot support a full methods-and-controls appraisal from the available coverage.

Funding, Application Dates and Living Costs

The MIBTP application page lists a 23 November 2026 deadline for October 2027 entry; it does not state the opening date, closing time or time zone, and current acceptance status is unconfirmed. Standard interviews are planned for late January/early February 2027, with final dates to follow. The current directory identifies the exact topic. The FAQ nevertheless says awards depend on candidate strength and that a shortlisted applicant may have to select another topic if the preferred project is unavailable.

Warwick explicitly covers the difference between UK and international tuition for successful international MIBTP awardees; UKRI supplies home-rate fees and maintenance. International awards may form up to 30% of the overall programme, not a 30% individual success rate. This is an affirmative funded international route, subject to selection, rather than automatic funding attached to every academic offer.

The published programme support includes a tax-free £21,805 annual stipend for 2026–27, training/conference travel, consumables and laptop use; the UKRI standard supports that dated figure. The 2027–28 rate and individual paid months remain unconfirmed. Four-year training is not a substitute for a personal award letter. Research budgets are not additional unrestricted income. The application form's £5,258 funding-source field is administrative information, not the maintenance stipend.

The official international cost page is explicitly based on October 2025 information. It states £1,171 monthly but also £13,632 for twelve months; multiplying the former gives £14,052, so the displayed amounts are inconsistent. Neither is a confirmed 2027 doctoral budget. The student-funding breakdown also references a 2025 survey. Campus rent includes certain bills; off-campus rent may leave utilities, internet and commuting additional. Full-year occupancy and holiday gaps matter.

An analyst planning range of £1,300–1,700 monthly, or £15,600–20,400 annually, leaves £1,405–6,205 against the older stipend before excluded upfront expenses. It is a scenario, not a measured expenditure, family budget or savings promise. The government fee references used here are £558 for a Student visa and £776 per visa year for student IHS, from official visa and health-surcharge guidance. Personal duration and future fee changes affect the bill. Relocation, deposits and documentation need separate cash; obtain explicit cross-university travel and temporary-housing reimbursement terms.

Entry Conditions, Mainland-China Eligibility and Preparation

The MIBTP academic rules accept a relevant first or high 2:1. A low 2:1 plus a relevant Master's, or at least three years of postgraduate experience, may qualify; a 2:2 or below normally requires exceptional additional experience. Overseas qualifications are explicitly considered for equivalence. The FAQ welcomes backgrounds including Psychology, Mathematics, Physics and Computer Science as well as experimental biosciences. It no longer imposes an A-level maths requirement, but expects preparation for intensive quantitative training.

The individual project does not publish additional verified numerical grades, mandatory programming certificates, an EEG qualification or a fixed number of papers. Psychology/cognitive science, statistics, computational neuroscience or related quantitative experience can support research match. Clear evidence of modelling and human-data interpretation is more useful than claiming every advertised technique has already been mastered. Those seeking only theoretical work, no participant contact, or a guaranteed clinical qualification should clarify the project's actual empirical commitments.

Warwick's official China page expressly welcomes applicants through research degrees and provides a Student-visa support route. Combined with MIBTP foreign-degree equivalence, international awards and Warwick fee-gap coverage, this provides positive evidence of in-principle eligibility for a domestic Chinese applicant. It does not establish automatic individual recognition. Warwick's detailed mainland-China postgraduate rules explicitly cover taught and research degrees and say only listed Chinese institutions are considered. UK 2:1 guides vary by institution, including 70%, 75%, 80%, and another group requiring 88% or above for individual consideration. These are not a uniform MIBTP high-2:1 threshold: check the actual institution and programme assessment. The undergraduate international-equivalence table is not a universal Chinese PhD admissions rule.

The current English policy assigns courses to bands A–D; the exact P-C1PBB band was not confirmed. The School of Life Sciences' published PhD guidance gives IELTS 6.5 overall with each component at least 6.0, while directing funded-DTP applicants to their specific guidance. This is a relevant school baseline; the precise P-C1PBB offer condition and English-band assignment still require confirmation. Tests must be within two years and one month of course start. An entirely English-medium degree in a UKVI-recognised English-speaking country, completed within two years and four months, may satisfy the condition at departmental discretion. An English-medium degree in China is not automatically covered by that statement. The FAQ permits later English evidence but retains an old October 2026 entry line, so obtain the 2027 requirement and individual exemption decision.

Applications use one Warwick P-C1PBB form, with one proposed supervisor and the full advertised title in the Research Proposal field, plus the separate ED&I form. A research-proposal upload is not an MIBTP requirement and may be replaced with a CV or statement. The FAQ asks for references, CV and cover letter; at least one referee must be academic, with a professional second referee permitted where necessary. Confirm upload slots and submit early because referees are emailed only after submission. Application instructions; supporting documents.

The ATAS rules do not list China as nationality exempt, but whether this specific course needs ATAS depends on the applicable subject/course code and offer. Do not infer an automatic ban or clearance from the topic. Academic admission, funding, any required approval, CAS and Student visa are separate conditions.

Decision and Questions to Resolve

The opportunity is Strong but verify fit. Its distinctive strength is the attempt to constrain a cognitive process with both behavioural and neural observations, supported by supervisors with complementary modelling and measurement expertise. Its main scientific risk is non-identification: plausible statistical correspondence may remain compatible with several algorithms or measurement models. The institutional and funding route is more concrete than the initial advertisement, while personal award terms and empirical resources remain consequential unknowns.

Before applying, contact Sanborn with the full topic title and a concise account of relevant modelling or human-data work. Confirm individual overseas-degree equivalence, the P-C1PBB English band and application materials. Follow the one-form/ED&I route and allow referees time before 23 November 2026. Verify whether the exact topic remains available; do not treat the current listing as a reserved place.

At interview, ask how Sanborn, Chater and Zhigalov divide regular supervision, behavioural work, neural acquisition and analysis. Which measurement site, access permissions, participant resources and technical support are committed? What does think-aloud evidence represent, how will reactivity and non-verbal processing be treated, and which competing observation or algorithm models can the available data distinguish? Discuss ethics approval, consent, data stewardship and plausible delays, without assuming a particular device or clinical sample. Ask what useful thesis questions remain if the anticipated behavioural-neural link is weak.

Before accepting, obtain the individual award's months, 2027 stipend, full international fee treatment and completion arrangements. Resolve formal degree/course and progression requirements, the 2027 training calendar, technical-visit duration and Aston-related travel/housing costs. Confirm equipment and data access rather than borrowing entitlements from earlier papers. Use a twelve-month housing quotation, and budget visa/IHS and other upfront costs separately.

Important limits remain: exact English band, individual Chinese equivalence, closing time, personal funded period, fixed empirical tasks and neuroimaging platform are not established here. Two of the four selected original papers were available only at primary abstract/theoretical-introduction level. These gaps narrow the assessment; they do not warrant inventing either a proven brain mechanism or a nationality restriction.

Evidence and Sources

Project, institution, eligibility and training

• Official Sanborn project, three supervisors, registration and supervision
• Current Neuroscience and Behaviour directory lists exact Sanborn project
• Sampling group official academic site
• Warwick Psychology Behavioural Science Group
• Alexander Zhigalov current Aston profile
• MIBTP 2027 application and institution-specific overseas support
• MIBTP applicant FAQ
• MIBTP training outline (headed 2025)
• QS 2027 Warwick overall ranking
• Warwick mainland China international admissions route
• Warwick postgraduate English bands
• Warwick mainland-China postgraduate institution requirements
• Warwick Life Sciences published PhD application and English baseline

Recent original research

• Zhu et al. 2022 Understanding the structure of cognitive noise
• Spicer et al. 2024 How do people predict a random walk
• Zhu et al. 2024 Autocorrelated Bayesian Sampler original institutional abstract
• Zhigalov and Jensen 2023 perceptual echoes original abstract

Funding, costs and immigration

• Warwick international living-cost guidance
• Warwick student-funding living/course-cost breakdown
• UK Student visa official route
• UK immigration health surcharge
• ATAS current nationalities and subject rules
• UKRI training grant current stipend