Admission

At least 8 PhD studentships for A.Y. 2023/24 will be available at the University of Bologna for candidates interested in any area of Statistical Sciences (start of activities: November 1st, 2023). Two of the eight studentships available at the time of this post are funded by NRRP ex D.M. 117/23; three other positions are funded by NRRP ex D.M. 118/23; they are all linked to the development of the PhD projects illustrated below.

All available positions will be covered by a scholarship. The award will be for three years, subject to satisfactory progress.

Special Topics Scholarships

Grant ex D.M. 118/23: SPATIAL MODELLING APPROACHES TO CLIMATE FINANCE AND SUSTAINABLE TRANSITION GOALS
PI: Silvia Romagnoli

The main topic of the grant is in line with the objectives of supporting the digital and ecological transition of the economic system and the administrative apparatus, with an expected impact on effectiveness and efficiency. Environmental impact assessment is at the centre of proper management and adaptation/mitigation policies necessary to achieve the objectives set by regulators. It is precisely in this context that the use of innovative strategies is linked to methodological and data innovation that can allow a distinction of the spatial location of the analysis and prepare targeted actions to highlight the major criticalities. Satellite data allows for objective and geographically localized analysis that promotes, on one hand, the valorization of the resource in question, and on the other hand, the determination of local risks related to natural variables and climate change, with an impact on the entire economic pattern, from the productive sector to the financial and the administrative one. The identification of risk-hedging solutions is among the objectives of mitigation and adaptation in which the spatial dimension plays a fundamental role given the nature of the problem itself.

References
- G. Christakos, “Spatiotemporal Random Fields: Theory and Application”, Elsevier, 2017 (second edition)
- F.E. Benth-J.S. Benth, “Modeling and Pricing in Financial Markets for Weather Derivatives”, World Scientific, 2013

Grant ex D.M. 118/23: MODERN STATISTICAL LEARNING FOR COMPLEX DATA IN HEALTH-RELATED SCIENCES
PI: Silvia Cagnone

In the field of health-related sciences, it has become increasingly common to deal with high-dimensional data. Reducing the dimensionality of such data is a challenging task in the statistical machine learning framework. Latent variable models (LVMs) provide a probabilistic solution for dealing with high-dimensional data by performing dimensionality reduction of correlated multivariate data to lower-dimensional latent variables. This project aims to explore inference tools for LVMs in complex data.

References
- Bianconcini, S., Cagnone, S., Rizopoulos, D. (2017) Approximate likelihood inference in generalized linear latent variable models based on the dimension-wise quadrature, Electronic Journal of Statistics, 11(2), pp. 4404–4423.
- Fan Dai, Somak Dutta & Ranjan Maitra A (2020) Matrix-Free Likelihood Method for Exploratory Factor Analysis of High-Dimensional Gaussian Data, Journal of Computational and Graphical Statistics, 29 (3), pp. 675–680.

Grant ex D.M. 118/23: STATISTICAL APPROACHES FOR DIGITAL HEALTH PREVENTION
PI: Saverio Ranciati

The rapid shift in the global health structure, mostly driven by an ever-ageing population demographics and new challenges brought forth by worldwide emergencies such as pandemics and climate change, has escalated the need for an efficient, ready, advanced, and sustainable approach to health. Due to the scale of these processes and their associated costs, in a wide socio-economic sense, digital prevention has become one of the prime candidates for a new paradigm to adopt to tackle such issues. Digital prevention refers to the landscape where data and information live, which is comprised of measurements through multiple devices, both born-as-medical or others, data recorded in public registries belonging to national health systems as well as health data belonging to the private sector. Given the digitalized nature of this health ecosystem and how complex the data living in it are, proper methodologies to analyze the information are vital, to (i) extract meaningful information; (ii) assess valid quantitative indicators for policymakers, stakeholders and, more importantly, the general population; (iii) improve all the related processes that are connected to health prevention, for example preventing via diagnostics life-threatening diseases, or acting on potential risk factors.
The project aims to understand, first, the landscape of statistical methods employed as state-of-the-art for analyzing digital health data, and secondly contribute to expanding or complementing these tools to improve their performances. The expected outcomes of the project will relate to: improving how and how much public administrations have to collect data such as personal health records, how to use them as information for statistical procedures and general data analysis strategies; aiding practitioners and researchers in defining protocols for digital health studies.

References
- Alami, H., Gagnon, M. P., & Fortin, J. P. (2017). Digital health and the challenge of health systems transformation. Mhealth, 3.
- Stark, A. L., Geukes, C., & Dockweiler, C. (2022). Digital health promotion and prevention in settings: Scoping review. Journal of Medical Internet Research, 24

Grant ex D.M. 117/23: STATISTICAL MODELLING FOR INVESTIGATING OUTDOOR AIR POLLUTION EFFECTS ON HEALTH
PI: Massimo Ventrucci, Andrea Ranzi (aranzi@arpae.it)

The PhD scholarship is sponsored by a PNRR project called “Aria outdoor e salute: un atlante integrato a supporto delle decisioni e della ricerca” with Regione Emilia-Romagna as principal investigator. The project main goal is to build territorial atlas mapping air pollution and disease risks at national level. The PhD project in Statistics has two aims. The first is to study and implement statistical models for disease mapping under the Bayesian framework which has a rich literature in environmental-spatial epidemiology (Lawson, 2018). The second point regards the study and implementation of ecological regression models for space-time data to investigate air pollution effects on health, both focusing on short-term and long-term effects; one further goal here is to consolidate results currently available in the literature on conventional pollutants (Whaley et al., 2021). From the methodological viewpoint, the class of models we will work with are Generalized linear mixed models (GLMM) for data aggregated over spatial units and time. The challenge will be to apply the GLMM framework to the air pollution and health data produced by the atlas, which will provide a platform to analyze impact of pollutants on several health outcomes at national level.

References
- Lawson AB. Bayesian disease mapping: hierarchical modeling in spatial epidemiology. 3rd ed. Boca Raton, FL: Chapman and Hall/CRC, 2018.
- Whaley P, Nieuwenhuijsen M, Burns J, editors (2021). Update of the WHO global air quality guidelines: systematic reviews. Environ Int. 142 (special issue)

Grant ex D.M. 117/23: MULTICENTER CLINICAL STUDY BASED ON THE USE OF VIRTUAL REALITY TOOLS AND 3D INTERACTION
PI: Monica Chiogna, Elena Toschi (elena.toschi@iss.it)

Virtual Reality (VR), which is increasingly being explored as a cost-effective digital therapeutic, is a technology that simulates a real environment. The recent literature suggests that VR-based interventions are able to improve symptoms of anxiety and depression and have been successfully used to treat acute and periinterventional pain in adults and children (Schrempf MC et al Sci Rep, 2022). The effectiveness of VR interventions has also been explored as a therapeutic tool in rehabilitation research (Laver KE et al Cochrane Database Syst Rev, 2017; Lei C et al PLoS One, 2019). Based on this, the main objective of the project, sponsored by UniBo (PNRR funds) in conjunction with Istituto Superiore di Sanità (ISS), is the elaboration of a multicenter randomized clinical trial investigating the feasibility and clinical effects of a personalized VR intervention and 3D interaction focused to improve the quality of life in patients affected by neurodegenerative or cancer disease. A study protocol will be designed with particular attention to: i) the assessment of the feasibility and clinical outcomes; ii) randomization plan; iii) sample size estimation; iv) statistical analysis plan.
The main educational objective of the project is to train a professional figure with advanced biostatistics expertise useful for a future job placement into public or private clinical health research institutions or Contract Research Organizations (CRO)/Clinical Trial Centres (CTC).

References
- Schrempf MC, et al. A randomised pilot trial of virtual reality-based relaxation for enhancement of perioperative well-being, mood and quality of life. Sci Rep. 2022 Jul 14;12(1):12067. doi: 10.1038/s41598-022-16270-8.
- Laver KE, et al. Virtual reality for stroke rehabilitation. Cochrane Database Syst Rev. 2017 Nov 20;11(11):CD008349. doi: 10.1002/14651858.CD008349.pub4.
- Lei C, et al. Effects of virtual reality rehabilitation training on gait and balance in patients with Parkinson's disease: A systematic review. PLoS One. 2019 Nov 7;14(11):e0224819. doi: 10.1371/journal.pone.0224819. eCollection 2019

Applicants may be of any age or citizenship, and must have a 2nd cycle degree or a single cycle degree from an Italian university or an equivalent qualification from other countries of at least four years' duration.

For courses that will start in the next academic year, the call for applications will be published at the page https://www.unibo.it/[…]/how-to-apply-phd-programme. The application must be submitted following the online procedure that will be linked in the call.

Applicants can also apply to the selection for the admission to the International PhD College of the Institute of Advanced Studies (ISA). See the page http://www.isa.unibo.it/en/activities/PhD_ISA for details. ISA PhD fellow will be offered free accommodation.

Admission is governed by a competition that allocates scholarships on the basis of comparative merit. Comparative assessment is made by the Boards of Examiners, appointed according to the University regulations, that evaluates qualifications and interviews candidates. After the selection process has been completed, the Board of Examiners draws up a merit list. Scholarships are allocated according to the ranking established in the merit list.

Qualifications include: undergraduate and/or master degrees transcripts of records, curriculum, scientific publications, GRE and TOEFL (if available), awards, grants, and one letter of recommendation.

The interview is meant to allow the committee to determine if the candidate is a good fit and if he/she has the motivation and drive to complete a doctorate. Moreover, it covers the technical background of each candidate. Technical questions are on the level, for instance, of:
- Exercise 7.6, 7.8, 8.3, 8,7 Casella Berger
- Exercise 7.4, 7.21, 8.2, 9.1 Mood, Graybill, Boes
Candidates are allowed to work on their answers with pencil and paper. They should then explain their answers to the examiners and answer additional follow-up questions.

In the event of positions linked to special topics, questions linked to the topic of interest might also be raised by the examiners on top of those above described. Candidates are invited to carefully read the topics, and, in case of interest, contact the PI's before the interview takes place.

Suggested readings for the interview:
- Casella, G., & Berger, R. L. (2002). Statistical inference (Vol. 2). Pacific Grove, CA: Duxbury.
- Mood A. Graybill F.A., Boes D.C. (1974): Introduction to the Theory of Statistics, McGraw Hill College.

The interview takes place on a date announced after the results of the evaluation of qualifications.

• Phd interview syllabus

  [ .pdf 94Kb ]

For more information, contact phd.stat@unibo.it