Psychology and Computational Modelling PhD: Can Emotion Improve Animal Decisions?

Psychology and Computational Modelling PhD: Can Emotion Improve Animal Decisions?

Quick Decision

This computational Psychology PhD tests when emotion-guided decision systems could improve biological fitness. The advertised work uses evolutionary modelling and artificial neural networks to study examples such as foraging under danger and mate finding under competition, then compares those systems with an unemotional Bayesian benchmark. The project is modelling-led; it does not assume that animals consciously experience states in a human-like way. It suits applicants who can bring quantitative or programming skills to questions in animal behaviour, evolution or psychology.

The main design question is whether the comparison is fair: how will “emotion” be represented, what information and costs will each system receive, and how will fitness be measured? The project does not yet specify its network architecture, learning rules, chosen scenarios or software. These choices should be discussed before applying; do not treat particular modelling extensions or empirical work as promised deliverables.

The Exeter award is for September 2027 entry and offers up to 3.5 years of funding, tuition support and a £5,000 research-training grant. The official pages conflict on the reference deadline (8 January versus 15 February 2027) and label £21,805 as both a 2026/27 and 2027/28 stipend rate. Plan for the earlier referee date and ask admissions to confirm the 2027/28 amount in writing.

Research Problem and Doctoral Work

What Is Unsolved and Why It Matters

The project asks why evolution might favour a decision system guided by emotion or mood instead of a direct stimulus-to-action rule. Its working idea is that an internal state can summarize prior experience and thereby inform decisions in environments that change over time. For example, physiological energy state can encode something about past food availability without requiring a fresh probabilistic calculation for every choice.

Higginson and colleagues provide a close theoretical precedent: in a foraging model with temporal autocorrelation, a physiological state-based strategy can approach Bayesian performance under some conditions. That result motivates this project, but does not show that all emotions are adaptive, that an animal consciously experiences the state, or that one physiological variable stands in for every emotion. Recent animal-emotion research emphasizes operational definitions and the limits of inferring an inner state from behaviour. The project is best understood as testing an adaptive decision mechanism, not as assuming human-like subjective feelings in non-human animals.

Work, Methods, and Potential Outputs

The project description advertises three linked tasks:

1. Build and evolve decision systems using evolutionary modelling and ANNs. It does not specify the neural-network architecture, learning rule, which components evolve, programming language or software stack.
2. Test systems in examples such as foraging under danger and finding mates under competition. The chosen scenarios, environmental dynamics, parameter values and species assumptions remain open.
3. Compare emotion-based strategies’ biological fitness with an unemotional Bayesian system that calculates an optimal decision from current information. The core design issue is to define what information and costs each system receives and how “optimal” and “fitness” are measured.

The description says the work may help explain some observed irrational or maladaptive behaviour. This is a possible implication, not a promised result. The project is described as modelling-led and does not commit to field experiments, animal testing or a particular dataset. Exploring how performance changes with environmental predictability, energetic or informational costs, or model misspecification could be useful, but these are ideas to discuss with the supervisors rather than confirmed thesis deliverables.

The doctorate is computationally plausible, but its central claims will depend on a fair comparator. At interview, ask how emotion is represented in the network, whether both systems receive equivalent information, and what would count as a meaningful fitness advantage. Those choices determine whether the comparison isolates decision mechanism or simply compares systems with different information and computational resources.

School, Programme and Field Context

The University of Exeter is the named institution. The project PDF places the work in its Department of Psychology, Washington Singer Laboratories, Streatham Campus. The 2027 funding round and project directory describe an Exeter-based, project-led award; no second host or degree-awarding institution is listed. Exeter’s Psychology research-degree page describes the usual MPhil/PhD route: students apply for an MPhil and normally upgrade to PhD by the end of year one, with research training and supervisory support. Confirm whether this is the registration route for this award and how its milestones fit the 3.5-year funding period.

For subject-level context, the department reports 87th in the 2026 Times Higher Education World University Rankings by Subject and 82nd in the 2026 QS World University Subject Rankings for Psychology. These are department subject rankings, not rankings of this project or its supervisors. In REF 2021, Exeter’s Psychology, Psychiatry and Neuroscience unit (UoA 4) received a profile of 53% 4-star and 39% 3-star activity. REF evaluates submitted research outputs, impact and environment across the unit; it cannot establish the resources available to this doctorate.

The Centre for Research in Animal Behaviour studies behavioural mechanisms and function in humans and other animals. Its page lists Tim Fawcett and reports weekly group meetings and a term-time seminar series, making it a relevant intellectual setting. Public information does not specify project-specific compute resources, software, training modules, teaching expectations, or any field component. Confirm those details with the team.

Supervisor and Lab Assessment

Supervisor Research Directions

The project names Tim Fawcett as lead and Danny Williamson as an additional supervisor. Fawcett is listed in Exeter’s Centre for Research in Animal Behaviour. Williamson’s official profile lists decision theory, optimisation, machine learning, Bayesian inference and statistical modelling—relevant expertise for the proposed comparator. Public information does not say how the two supervisors will divide the work. Ask who will guide evolutionary theory, network implementation and behavioural interpretation, and what regular supervision arrangements are planned.

Recent Important Publications

• Olivier, Fawcett and colleagues (2025), “Cooperative breeding and the selection for information sharing among groupmates.” Their evolutionary model studies how imperfect information about subordinate quality changes reproductive concessions and incentives to acquire or conceal information. It illustrates how information assumptions can affect evolutionary outcomes, but is not an emotion-based ANN model.
• Burnham et al. (2026), “Pole position: social influences on discounting behaviour in wild grey squirrels.” A naturalistic task found squirrel choices between easier/closer and preferred/farther food varied with social rank and conspecific presence. It shows why ecological and social context can matter for decision-making, including the advertised mate-competition example. It is not an emotion manipulation or the proposed model.
• Wilde et al. (2025), “Biomimetic robots reveal flexible adjustment of sexual signalling in a wild invertebrate.” Manipulating apparent rivals with robots showed flexible fiddler-crab sexual signalling. This is relevant empirical context for mating under competition, not a forecast of the PhD’s method.
• Langford et al. (2024), “Bio-logging reveals heritable patterns of natural behaviours in sheep.” Analysis of collar acceleration from 84 ewes over 28 days identified repeatable behaviour profiles, one heritable component and environmental associations. It signals quantitative animal-behaviour research in the wider group, but is observational rather than the advertised modelling approach.

Taken together, these publications support a research environment spanning evolutionary modelling and ecological animal-decision research. They do not establish a fixed PhD architecture, guaranteed empirical component, or access to specific datasets or computing facilities.

Lab, Resources, and Collaboration Environment

The Centre’s public page describes weekly group meetings and a term-time seminar series and lists research students and technical staff. This suggests regular exchange across animal-behaviour research. It does not document GPU access, software licences, dedicated modelling support, data infrastructure or travel/conference funding. The current project PDF names two Exeter supervisors and no external partner or joint degree. Ask directly about compute, training and supervision.

Funding, Application, and Living Cost

Award Structure and Eligibility

This project is attached to Exeter’s September 2027 Psychology Doctoral Landscape Award. The associated award page describes an index-linked stipend for up to 3.5 years full-time, home tuition-fee coverage and a £5,000 Research Training Support Grant over the award. It states that Exeter covers the home/international fee difference for successful international awardees. The central page says international fee-status students may occupy up to 30% of the cohort. Award selection and eligibility still apply; international students must budget for visa, immigration health surcharge and moving costs.

For 2027/28, the general academic threshold is normally a First or Upper Second UK honours degree, or overseas equivalent, in an appropriate science, engineering or technology subject. A lower second may be considered alongside a Master’s or significant relevant non-academic experience. The project-specific PDF separately asks for a “good BSc or MSc” in Biology, Zoology, Psychology, Mathematics, Computer Science, Data Science or another strongly quantitative subject. “Good” is not numerically defined, so confirm the project-specific standard instead of assuming that the general scheme threshold settles it.

The published £21,805 stipend has inconsistent year labels: the main award text calls it the current 2026/27 rate, while the award-page summary labels the same amount as the 2027/28 rate. The central page also presents it as the current 2026/27 figure. Treat it as a reference only and ask for the indexed 2027/28 amount and payment schedule. International fee coverage does not remove visa, health-surcharge, relocation or other moving costs.

Application Route, Documents, and Dates

Apply through the link in this project’s PDF so the application reaches the right panel. The stated documents are a personal statement about academic interests, prior research and reasons for choosing the project; CV; complete transcript(s), including an interim transcript if still studying; two references; and English-language evidence where applicable. Documents must be in English, and qualifications in another language need certified translations.

Both official pages state the application deadline as midday on Tuesday 8 December 2026. The reference dates conflict: the project-associated award page says 8 January 2027, while the central scheme page says 15 February 2027. Missing documents may prevent shortlisting, so plan for the earlier date and request written confirmation. The central page also has one stale paragraph referring to interviews in February 2026 and outcomes in March 2026, although other current information gives 2027. The award page says general interviews are likely between 15 and 24 February 2027. Confirm the current interview schedule rather than relying on the stale paragraph.

Exeter Cost and Stipend Buying Power

Exeter estimates £1,300–£2,000 per month for basic expenses for a single student in Exeter or Cornwall, using average local spending in April 2026. The estimate includes accommodation, food, transport, utilities and other basics; the university advises budgeting for arrival costs such as bedding or kitchen equipment.

At the published £21,805 reference amount, the stipend equals about £1,817 per month before personal deductions: about £517 above the low estimate and £183 below the high estimate. This is only a planning comparison. The final 2027/28 amount, personal circumstances, travel, initial costs and (for international applicants) visa and health-surcharge costs may change it. As a housing example, the university lists Cook Mews at £5,470.74 for a 44-week 2026/27 contract and Llewellyn Mews at £6,343.92 for 51 weeks; both are described as roughly a 15-minute walk to Streatham. These are not 2027/28 availability or price guarantees.

Applicant Fit and Preparation

Required Background and Skills

The project asks for a “good BSc or MSc” in a named biological, psychological or quantitative field. Separately, the award normally expects a First/2:1 equivalent in an appropriate science, engineering or technology subject, with the possible lower-second route noted above. The project’s “good” wording is not defined, so ask how the team evaluates grades and relevant experience.

A strong application will likely need to show quantitative reasoning and capacity to implement or learn computational models, alongside an ability to interpret biological assumptions about fitness, internal state and environmental change. The advert does not specify a programming language, neural-network package or prior animal-emotion research requirement. English proficiency evidence applies where required.

Preferred Experience and Suggested Preparation

The project description does not provide a separate list of preferred experience. Useful preparation would include quantitative coursework, code or reproducible simulation examples, probability/Bayesian reasoning, and an example of interpreting model results in evolutionary or behavioural terms. These are suggestions for demonstrating readiness, not additional published entry conditions. Review the project-cited physiological-state model and prepare to discuss how a fair Bayesian comparison could be designed without presenting unconfirmed methods as settled.

UK career profiles describe data science, AI/ML and life-science research as possible adjacent pathways. The project may build transferable modelling and programming skills, but it does not promise an industry placement or AI-sector job. Graduates targeting those roles should also develop software practice, data handling and a portfolio relevant to the specific job.

Next Actions and Questions to Verify

Before Applying

• Ask Exeter whether references are due 8 January or 15 February 2027; use 8 January as the safe planning date until confirmed.
• Request written confirmation of the 2027/28 stipend and payment schedule.
• Clarify what “good BSc or MSc” means for this project and whether a separate research proposal is required.
• Apply through the project-PDF link and confirm the current interview timetable.

At Interview

• How will emotion or mood be represented in the ANN, and which components evolve?
• Which scenario will anchor the thesis, and how will environmental change, danger, competition and costs be varied?
• What information and costs does the Bayesian comparator receive, and how are optimal decisions and biological fitness measured?
• Who leads theory, implementation and behavioural interpretation, and what coding languages, compute resources and training are available?
• Is any empirical/field component expected, and what data or facilities would it use?

Before Accepting an Offer

• Confirm MPhil/PhD registration, upgrade milestones and training requirements against the funded period.
• For an international offer, confirm full-fee coverage for every funded year and budget separately for visa, health surcharge and relocation.
• Confirm supervision schedule, project scope, RTSG rules and any travel/conference costs.

Evidence and Sources

Core Institutional and Programme Sources

• Official project description: The emotional basis of behaviour
• Psychology DLA award and application information
• EPSRC DLA eligibility and central timetable
• 2027 EPSRC DLA project directory
• Psychology MPhil/PhD programme
• Department of Psychology rankings and environment
• REF 2021 Psychology, Psychiatry and Neuroscience profile
• Centre for Research in Animal Behaviour

Supervisor and Publication Sources

• Tim Fawcett academic profile
• Danny Williamson Bayesian-inference profile
• Olivier et al. (2025), information sharing in cooperative breeding
• Burnham et al. (2026), social influences on squirrel discounting
• Wilde et al. (2025), robotic rivals and sexual signalling
• Langford et al. (2024), bio-logging sheep behaviour

Field and Methods Sources

• Goldshtein and Yovel (2024), sensors and animal decision-making
• Bell (2024), evolution of decision mechanisms
• Paul and Mendl (2025), challenges in animal-emotion research
• Review of animal-emotion mechanisms (2025)
• Procenko et al. (2024), judgement bias in stressed bees
• Baciadonna and Nityananda (2026), bumblebee judgement-bias consistency
• Higginson et al. (2018), physiological state and Bayesian learning

Funding, Cost and Career Sources

• UK AI Labour Market Survey 2025
• National Careers Service: data scientist
• Prospects: life-sciences research scientist
• Exeter student budgeting estimate
• Cook and Llewellyn Mews accommodation prices