Our regulation of sleep, cognition, and metabolic state is driven by a central clock, which is in part, trained by environmental triggers. An understanding of the circadian regulation of mood is vital for coping with day-to-day needs. But in-depth understanding has been hampered by the need for large datasets and objective reporting. Research usually relies on questionnaires but these can be subject to recall bias.
One way to avoid recall bias is to get an objective snapshot of mood at the time it is being felt. Our social media trends offer researchers a sea of data and the ability to assess other people’s moods objectively, although false reporting does need to be factored in as a potential barrier.
The EU’s THINKBIG project’s researchers used a massive dataset of over 800 million Twitter messages collected over 4 years in the United Kingdom. The team surveyed anonymised tweets from the 54 largest towns and cities in the UK every 10 minutes to investigate how social media can provide insight into mental wellbeing throughout the day and year.
They extracted robust signals of the changes that happened during the course of the day in the collective expression of emotions and fatigue. The team used methods of statistical analysis and Fourier analysis
to identify periodic structures, extrema and change-points along with comparing the stability of these events across seasons and weekends.
Researchers were careful to take into account greetings and holiday messages and extracted key vocabulary that could have skewed the results. They ignored any post containing the word ‘happy’, ‘merry’, ‘good’, ‘lovely’, ‘nice’, ‘great’, or ‘wonderful’ followed by ‘Christmas’, ‘Halloween’, ‘Valentine’, ‘Easter’, ‘New Year’, ‘Mothers day’, ‘Fathers day’, and their variants.
Their paper Circadian Mood Variations in Twitter Content
published in the journal ‘Brain and Neuroscience Advances’, reveals strong, but different, circadian patterns for positive and negative moods. They write, ‘(…) the cycles of fatigue and anger appear remarkably stable across seasons and weekend/weekday boundaries. Positive mood and sadness interact more in response to these changing conditions. Anger and, to a lower extent, fatigue show a pattern that inversely mirrors the known circadian variation of plasma cortisol concentrations. Most quantities show a strong inflexion in the morning.’
Building on past research
This is not the first time Twitter has been used as a data source. This study expands upon those earlier studies, using a larger dataset, collected over a longer time interval from a narrower geographic area. THINKBIG also used refined methods of analysis and robust mood indicators.
They found circadian patterns that are compatible with previous observations, confirming that positive and negative emotions are periodic quantities independent of each other. ‘We believe this study is the first to decompose the spectrum of negative emotions into anger and sadness and to compare them with fatigue.’
The team did uncover evidence that emotionality maybe closely associated with cortisol levels and even to circadian cortisol release. In the light of this, the results they found showing anger and, to a lower extent fatigue, inversely mirrors the known circadian variation of plasma cortisol concentrations is quite striking. They acknowledge this association must be treated with caution as we there is no evidence for causality.
THINKBIG (Patterns in Big Data: Methods, Applications and Implications) secured access to the collection of all UK newspapers from the past 200 years, which they set out to analyse along with working closely with colleagues from a variety of disciplines to expand their current work on social media mining. The project intends to have an impact on the social sciences, the general public and the lawmakers, along with their key field of engineering.
For more information, please see: Project website