Teija Kujala, Markus Jokela and Aapo Hyvärinen are the most cited researchers in the field of cognitive sciences at the University of Helsinki.
Highly cited researchers: cognitive sciences
Teija Kujala (b. 1964), professor in psychology, has conducted pioneering research on treating dyslexia as well as on the early development and plasticity of the brain in utero. Her studies have garnered a great deal of attention both in academia and in popular media around the world.
In 2001 Kujala and her colleagues found that dyslexic first-graders learned to read more effectively when they did exercises on a computer involving combining strings of blocks on the screen with melodies played through the speakers.
The researchers found that the children’s capacity to discriminate between speech sounds increased both in practical tests and in the brain as established by their mismatch negativity (MMN) response in an EEG.
MMN, a measurement originally created by Professor Risto Näätänen at the University of Helsinki and further developed by Kujala, measures the brain’s response to changes in the stimulus environment and can be used to study the ability of sleeping babies to differentiate between sounds, for example. MMN is widely used all over the world, and Kujala is much cited for her articles on the use of the method.
Kujala’s mission is to prevent serious difficulties in learning to read. In 2012, her research group found that pre-schoolers considered at risk for dyslexia improved their reading readiness with just three hours of practice with the Graphogame programme developed by Professor Heikki Lyytinen.
In 2013, Kujala’s research group shocked the world when they discovered that it is possible to begin brain training in utero.
In the test group, pregnant women played a CD recording of the nonword “tatata” to their bellies with periodical variations in the length and pitch of the vowel in the middle of the utterance.
During their first weeks of life, the infants born to the test-group women were tested for their MMN response to variation in the “tatata” nonword. The response in the test-group infants proved to be greater than in infants not exposed in utero to the material in question.
In the next study, researchers repeatedly played a recording of Twinkle Twinkle Little Star to fetuses in utero. At age four months, the brains of the infants in the test group responded to the melody more intensely than did infants in the control group.
Thus a direct study of infant brains established for the first time in the world that the brain can learn from its environment already in the womb and that in utero experiences leave long-term memories.
In 2014, Kujala and her colleagues launched an eight-year research project which tracks the linguistic development of 200 children from birth until school age. Using extensive batteries of brain response tests, the researchers will measure the children's sound discrimination ability and their linguistic development. The goal is to determine whether music-based linguistic training during the first six months of life can improve the children’s linguistic learning and whether it could serve as a tool to prevent dyslexia.
In her dissertation article in 1995, Kujala was among the first researchers to find that the visual cortex of blind people processes auditory stimuli.
Before focusing on dyslexia and infants, Kujala studied background noise extensively. She discovered that people working in noisy environments experienced long-term deterioration of their ability to discriminate between speech sounds, and that unexpected sounds were more likely to break their concentration.
In addition to leading her own research group, Kujala is co-director at the Cognitive Brain Research Unit together with Professor Mari Tervaniemi. Kujala's most cited studies involve a wide array of partners from European and American universities.
Kujala herself has near perfect pitch, but primarily uses it to listen to heavy metal.
Markus Jokela (b. 1979), associate professor in psychology, is an anomaly among the University's most-cited researchers in that he has gained high numbers of citations with zero-result studies, which proves, like academic mythbusters, that a presumed correlation just doesn’t exist.
Jokela combines large-scale follow-up studies from different fields of science into one statistic, with sample sizes in the tens or hundreds of thousands. The emergence of a correlation between individual factors (e.g., personality, intelligence, health behaviour) and population-level phenomena (e.g., birth rate, regional differences, migration) in such a large dataset gives the study high credibility, and confirms the connection.
In 2009, Jokela used data collected in the USA to establish that a combination of extrovert tendencies, low agreeableness and a high level of openness to new experiences increased the likelihood that an individual would migrate within the country. The result makes sense and seems obvious, but until that study, migration had not been examined academically from a perspective of individual psychology, as studies tended to focus instead on sociodemographic factors such as age, gender, education and wealth.
Combining such different data is relatively rare, but interest in this field of research is growing, which is also bringing citations to Jokela as a pioneer of new research topics.
Correlations reported in small-sample studies often fade when large amounts of data undergo meta-analysis. Publication bias, or the likelihood that studies “revealing” strong correlations are more likely to be published than zero-result studies, is a common problem in academia. In psychology, small samples have shown many astounding results, which other small-sample studies have successfully repeated.
Jokela is a fierce proponent of methodological integrity and the importance of zero-result studies.
Since 2013, Jokela has conducted many “individual participant” meta-analyses. Such studies test a theory by compiling several datasets which could answer the research question, and then the analysis is run through all of the data. This helps alleviate publication bias, as the researcher can use the original data.
In one of his first individual participant meta-analyses, Jokela established that high conscientiousness correlates with low mortality risk. However, the extensive data also showed that sociable, neurotic or agreeable tendencies had no impact on mortality. Researchers who had found such correlations in their own smaller studies naturally found these results, gleaned from a set of data from nearly 80,000 people, unwelcome.
Jokela has also established a link between high conscientiousness and low body-mass index, but found no evidence of links between any other personality traits and obesity in his data of 80,000 test subjects.
Jokela has also shown that personality type does not correlate with cancer.
In his recent study involving data from 600,000 individuals, Jokela found that a working week of more than 55 hours increases one’s risk for stroke by more than 33%.
Jokela is currently working on topics related to population-level mental health.
Jokela holds a doctorate in psychology from the University of Helsinki and in epidemiology from University College London. His involvement in helping other psychologists, public health researchers and social scientists in their statistical analyses has also boosted his citation figures.
Aapo Hyvärinen (b. 1970), professor of computer science, is known as a developer of revolutionary algorithms. He specialises in machine learning and independent component analysis, in particular ICA. In ICA, masses of data are searched for interesting original components.
In computer science, Hyvärinen is often cited for his algorithms themselves. Many of the citations also come from neurosciences, where Hyvärinen, or at least his algorithms, has helped filter the huge amounts of data acquired through brain imaging.
In 1995, Hyvärinen began to write his dissertation on ICA at the Helsinki University of Technology. During his first year of studies, Hyvärinen conceived an algorithm that could perform the required computations faster, but without compromising accuracy. A key insight was developing a fixed-point algorithm for analysing independent components.
This algorithm turned out to be better than the others that had been put forward, and it was widely adopted. It has been used to analyse groups of data in many disciplines throughout the 2000s. For example, images from space or genome maps have yielded interesting components. Hyvärinen’s innovation from his first year of postgraduate studies continues to be his most cited discovery.
After the dissertation, Hyvärinen has used ICA to study the brain in particular.
In 2000, he managed to use ICA to model the function of the complex cells in the primary visual cortex. Before this, ICA had only been used to model the simpler cells in the visual cortex. From 2007, Hyvärinen has been working on ICA algorithms to interpret brain imaging data while cooperating with researchers from Japan and elsewhere.
Current ICA algorithms can only identify linear dependencies. This means that the derived data can be expressed as a linear function describing independent components in the real world. Non-linear, more complicated functions are beyond the grasp of ICA.
Hyvärinen is now about to publish an ICA algorithm which can also discover non-linear dependencies. A central innovation was to include the temporal element: observations made at short intervals are usually more similar than ones made further apart in time. This offers a point of entry for identifying complex regularities.
German researchers published a similar discovery in 2003, but they were unable to create an exact mathematical theory based on it.
Throughout his career, Hyvärinen has benefitted from his ability to refine his insights into exact mathematical theories and, consequently, to functional computer algorithms which can seek regularities.