Jesus M. Cortes, Ikerbasque Professor at the Biocruces-Bizkaia Health Research Institute and Head of R&D at NeuronUP, explains in this interview the findings on the prediction of cognitive impairment with NeuronUP that have just been published in the prestigious British Journal of Neuropsychology.

What have been the major findings in this study?

The motivation of the study was to rigorously understand to what extent the different cognitive training materials developed by NeuronUP could be used to predict cognitive impairment one year after using NeuronUP.

Out of 203 different cognitive training materials studied, we have understood that some materials perform better than others. For example, materials for training processing speed, attention (in its three forms—selective, alternating, or sustained), and executive function had predictive power with area under the curve accuracies greater than 0.89, which are very high compared to previous studies.

Furthermore, in more specific pathologies, we found differences such as selective attention predicting cognitive impairment well after 1 year in Parkinson’s but not in Alzheimer’s, where most existing materials predict cognitive impairment moderately. We have also studied populations with multiple sclerosis or Down syndrome.

On the other hand, this study has allowed us to develop a data engineering infrastructure within NeuronUP, greasing the wheels of the whole machinery, and creating a specific data unit, where there is already dedicated personnel within this unit.

Why was the Journal of Neuropsychology chosen to publish these results?

The Journal of Neuropsychology is a publication of the British Psychological Society, one of the oldest and most respected psychology organizations in the world.

Founded in 1901 and with over 60,000 members, it is a professional organization for psychologists in the UK, organizing a wide range of activities and services for its members, including continuing education, research, and promoting psychology.

It is a strong advocate for promoting psychology for the public good and works closely with other organizations and government agencies to improve mental health and well-being in the UK and worldwide.

The British Society manages several prestigious scientific journals with a high impact factor, covering clinical psychology, neurodevelopmental, educational, health, social, and neuropsychology. It is, without a doubt, a forum of quality and established prestige for the neuropsychology professional.

On the other hand, our research utilizes disruptive technology in neuropsychology, based on Real World Data, and this has somewhat penalized us considerably. In particular, one of the reviewers was very demanding and rigid, and it was very challenging for us to publish our results in this journal. However, this was a risk we decided to take, as our strategy was that if a reputable journal in the community published our work, professionals and researchers would be more receptive to our methodology.

Why is Real World Data-based technology disruptive?

I would like to emphasize first what Real World Data (RWD) is and why it is different from typical methodology in clinical research. RWD refers to data collected outside the controlled environment of a clinical study, as in everyday clinical practice or traditional clinical research.

RWD mixes data from electronic health records, online platforms, surveys, health insurance, among others. Unlike data collected in controlled clinical studies, which have very little generalization power to populations other than those studied, RWD can provide a more precise and generalizable view of how a treatment or intervention behaves in the real population.

The downside is that RWD is highly heterogeneous and requires the use of more agile non-traditional technologies like machine learning or artificial intelligence.

RWD is a disruptive technology because it allows access to a large volume of data and a wide variety of patients, which enables a better characterization of populations and a better understanding of interactions between treatments and comorbidities.

It can also help identify new therapeutic indications and evaluate the safety and efficacy of treatments in a setting closer to real clinical practice.

On the other hand, it faces new challenges, such as increased complexity in data collection and cleaning, in data analysis itself, less control over confounding factors, and legislation not designed for RWD, with recurrent issues of data privacy and security.

How was cognitive impairment measured in your study?

Cognitive impairment in our study was defined based on population data in performance when completing NeuronUP cognitive training materials. Essentially, in cohorts of tens of thousands of participants, we can well define what is normal and what is impairment by simply defining percentiles in each patient’s scores relative to the population.

The performance of each participant was measured based on the NeuronUP Score, the score we use to measure users’ progress in NeuronUP. It is a index between 0 and 100, different for each participant, and calculated using a formula that combines correct answers, time spent on the test, and the difficulty level of the test. NeuronUP Score is a novel quantitative index that simplifies participant performance and facilitates comparisons in tracking, allowing the longitudinal data modeling of the same participant and obtaining precise individual performance trajectories.

How do you think your study could contribute to a better understanding of the causes of cognitive impairment?

This is a very ambitious question and we still don’t know how to answer it. To explain the causes of cognitive impairment, we should study the genetic and environmental factors related to this condition. We should also address the different forms of cognitive impairment, such as dementia, and the differences in the underlying causes of each type.

Moreover, the changes in the brain and in cognitive and neuropsychological function between individuals with and without cognitive impairment should be characterized, as well as their peculiarities in specific populations, such as older individuals or individuals with certain underlying medical conditions, to identify specific risk factors for this condition.

Furthermore, therapeutic interventions and their effectiveness in improving or preventing cognitive impairment should be characterized. While more research is needed in this area, our article using RWD and NeuronUP Score is a good starting point and motivates the need for further studies in the future using NeuronUP Score and specific cohorts to advance in this issue.

How do you believe your findings could be used to improve healthcare for people with cognitive impairment?

Quantitative studies to predict cognitive impairment with high accuracy in the general population up to 12 months before it occurs are very useful, for example, to identify individuals at risk of developing cognitive impairment in advance.

This would allow for early intervention to delay cognitive decline, design personalized prevention and treatment programs for individuals at higher risk of cognitive impairment, monitor the progress of individuals with cognitive impairment over time to assess the effectiveness of treatments, identify modifiable risk factors, and develop interventions to reduce the risk of cognitive impairment, or even assist doctors in making informed decisions about treating individuals with cognitive impairment.

What steps are being taken to carry out further research within the NeuronUP R&D Unit?

In this unit, we started working in 2018 with funding from the Center for Industrial Technological Development of Spain (CDTI), under the Ministry of Science, Innovation, and Universities, and from the Economic Development Agency of La Rioja (ADER), as well as with NeuronUP’s own internal funding.

During this time, we have managed to optimize many aspects of data collection, engineering, and monitoring, as well as quantification and visualization, of high value to the clinical professional.

Moreover, we have started collaborating with several leading research centers in the world (national and international) on different research projects, such as the detection of outliers and their monitoring in the general population, validation of NeuronUP Score through the use of standard neuropsychological tests, or the classification of trajectories of each participant according to their NeuronUP performance.

Without a doubt, in the coming years, we will be able to offer new functionalities based on the data for the use of the clinical professional.

If you want more information about the published study take a look at the following article: