Dr Alex Mielke

Alex Mielke

Lecturer in Psychology
Open Science Lead

School of Biological and Behavioural Sciences
Queen Mary University of London
www.qmul.ac.uk/sbbs/staff/alexander-mielke.html
Centre for Brain and Behaviour
ORCID Google Scholar X

Research

Primates, Cooperation, Communication, Sequences, Grooming, Play

Interests

Research Interests:
Primate Sociality and Cooperation

During my PhD, I studied sooty mangabeys and Western chimpanzees in the Tai National Park, Cote d'Ivoire, trying to device ways to compare their social systems. This has led to two broad interests: one, using naturally occurring social behaviour (grooming, grooming interventions, food sharing, spatial association) as decision making situations to quantify the complexity of cognitive processes in wild animals; the other, quantifying the social structures of animals (hierarchies, friendships, association networks) using long-term observational data.



Reproducibility and Replicability in Observational Studies

One result of the work on sociality is the realisation that there are way too many different ways to calculate basic aspects of animal social life from data, and that our data pipelines and current publishing formats do not ensure that two researchers, with the same data, derive the same insights about the animals they study. As most of comparative ethology is indirect (we read published papers and compare our results to theirs, even if they used a different method), we do not know whether reported differences between species are due to biological differences or because researchers used different methods to calculate dominance hierarchies, social networks etc. As a first step, we have started using simulation studies to quantify the impact measurement and sampling error have on different standard methods in animal sociality research, how the choice of different dominance or sociality measures affects results, and how different linear model specifications influence researcher interpretations.



Animal Play

Classically, studies of play have focused on partner choice - who plays with whom. In my work, I focus on the sequence of actions and reactions between players to test whether there are recognisable games and how predictable play is from the viewpoint of the individuals involved. As it is inherently more chaotic, and often faster, than 'serious' behaviours, play can be a measure for the most complex sequences an animal might encounter in their development. It is therefore a good indicator whether complex hierarchical sequence patterns, as we see them in human language, are evident in action sequences of animals - is there a 'grammar' of play, and how does it compare between species.



Non-verbal Communication - Gestures, Calls, and Facial Signals

I've been fortunate enough to work in groups that use different channels to understand non-verbal communication in humans and primates. I am currently involved in work on the syntax of chimpanzee gestural communication. I previously helped develop network-based approaches to quantify facial communication in humans and other species. In mangabeys, we showed the alarm calls follow predictions made by game theoretical models. Throughout, I have used of machine learning models for the study of animal communication systems (here and here for vocalisations; upcoming for gestural communication). I am working to prepare standard tools to analyse sequences of animal behaviour and communication to improve replicability and reproducibility of comparative research in this area.