Louis C H Fourie

Alzheimer’s disease and other forms of dementia (a collective name for brain syndromes which affect memory, thinking, behaviour and emotion) are a growing public health problem all over the world – with 10 million new cases worldwide every year or one new case every 2.6 seconds.

According to the World Alzheimer Report of 2020, there are more than 50 million people worldwide living with dementia in 2020, of which 60-80% are people with Alzheimer’s. This number doubles every 20 years and will reach 152 million in 2050. Unfortunately, 59.8% of people with dementia (29.83 million) are living in developing countries with a low or middle income. This will rise to 70.9% (107.94 million) by 2050. What exacerbates the situation is the economic impact of dementia, which is estimated at a global annual cost of about R14 trillion.

Research has shown that approximately three-quarters of people living with dementia have not received a formal diagnosis, especially in low- and middle-income countries. Since it is of critical importance to diagnose dementia as early as possible to enable appropriate and timely intervention, a cost-effective way of identifying dementia has become a high priority.

Paper-based dementia tests

MORE ON THIS

• Tech start-ups are ready for a national start-up act
• Creating smartphones for the 5th industrial revolution
• Telematics: A driving force for SA e-commerce

Until now a variety of paper-based tests have been used, such as the Community Screening Instrument for Dementia (CSID), often in combination with the Five Words Test, Animal Fluency, the Ten Word Delayed Recall Test, Stick Design Test, and Blessed Dementia Scale. Some other screening protocols that are used are the Ten Item Semi-Structured Home Interview (CHIF), the Mini-Mental State Exam (MMSE) or Folstein Test, The Addenbrookes’ Cognitive Examination 3 (ACE-3), the Montreal Cognitive Assessment (MoCA), and the Tygerberg Cognitive Battery (TCB). The screening phase is usually followed by clinicians applying the Diagnostic and Statistical Manual of mental disorders (DSM-5) criteria to confirm dementia diagnoses.

Innovative artificial intelligence (AI) test

Now two Cambridge University PhD graduates, through their company Cognetivity, came up with an innovative new smart phone app to help diagnose Alzheimer’s Disease and other forms of dementia, which takes only five minutes to execute and is more accurate than current paper-based tests.

The test uses explainable artificial intelligence (AI) to assess a person’s brain function by presenting them several black and white photos and requesting them to identify which ones contain an animal. The images are black and white so as not to disadvantage people who may be colour blind, as well as to remove any hidden clues present in colour, such as the specific colour of certain animals.

The AI-based Integrated Cognitive Assessment (ICA) test is based on humans’ strong reaction to animal stimuli, and the ability of a healthy brain to process images of animals in less than 200 milliseconds. Various mental disorders, specifically neuro-degenerative disorders, are phenotypically characterised by some degree of cognitive impairment. This rapid visual categorisation test engages brain areas affected in pre-symptomatic stages of Alzheimer’s such as the retina, visual cortex and motor cortex. It can detect subtle impairments in information processing speed, thus detecting early signs of the disease before the onset of memory symptoms.

The images appear very briefly for a split second only. Some images will clearly show an animal, while in others the presence will be less obvious or there will be no animal at all. The reason why animals are used is that research has proven that animals elicit strong reaction in people and thus provide a greater insight into a person’s brain activity. Images are used since it is not subject to linguistic or cultural biases of existing tests and can be used repeatedly to monitor development. Existing tests can be learned by subjects and therefore become less effective over time, according to the proponents of the use of images. The test is also not influenced by educational level.

A highly sensitive test for early detection

What makes the ICA app so valuable is that it gives an objective, highly sensitive measure of cognitive function, as well as an AI explanation of the model prediction. It can identify differences in the neural response speed of visual information processing long before the memory loss that current tests focus on, and thus allows the detection of dementia up to 15 years before the appearance of any symptoms. Although Alzheimer’s cannot be reversed, early detection does provide an opportunity to stop it through a variety of promising new drugs to treat early-stage Alzheimer’s such as aducanumab, lecanemab, donanemab and solanezumab (all monoclonal antibodies recruiting the immune system to remove beta-amyloid plaques in the brain), saracatinib (preventing destruction by reversing memory loss) and beta- and gamma-secretase inhibitors (blockers of beta-amyloid production).

Accuracy and adoption

A study described in a scientific paper currently under peer review, found the app to be 84.2% accurate at identifying people who are cognitively impaired, compared to 81.6% for the standard Montreal Cognitive Assessment (MoCA) test. In another trial, the ICA achieved 92% accuracy compared to 84% of the Addenbrookes’ Cognitive Examination (ACE) test. The developers expect the app to become even more accurate as the AI programme processes more data through machine learning.

In the United Kingdom, the Medicines and Healthcare Products Regulatory Agency (MHRA) has approved the inexpensive smart phone app and it is therefore already used in general practices and hospitals for the screening of cognitive impairment and dementia. According to the developers of the app, it may soon be used in homes on a smart phone to detect cognitive changes in ageing adults remotely and on a large scale. The app is also being used in the Tehran University of Medical Sciences in Iran since 2020, where it was found to be highly accurate, easy to understand even by poorly educated people, and easy to administer with automated scoring and AI to improve classification accuracy.

Until now, doctors had to mostly rely on paper-based assessments and expensive brain scans to support dementia diagnoses. The app test is cost-effective, accessible, simple to use and accurate in identifying people at a much earlier stage than current methods. Until now, it was also not possible to monitor patients with mild cognitive impairment due to time and cost. The efficiency, objectivity and ease-of-use of the app test could bring about a breakthrough in tackling a major health-care and economic problems like dementia.

Detection of multiple sclerosis

According to the researchers, an added benefit is that although the app has been developed for dementia, it could also be used to detect signs of multiple sclerosis (MS) long before any symptoms begin to appear. Cognitive impairment is common in MS patients, which means that the app could be used to detect signs of MS. Dr Masood Nabavi, from the Royal Institute for Stem Cell Biology and Technology, has recently published a study in the BMC Neurology Journal claiming that the test could distinguish between “cognitively normal” and “cognitively impaired” patients with an impressive 95% accuracy. The test could therefore be used as a marker for cognitive impairment in MS and to monitor the response of the patient to therapy.

According to the chief executive of Cognetivity, Dr Sina Habibi, the technology is “capable of revealing underlying physical damage to brain cells without the need for invasive measurement such as blood or spinal fluid sampling” and is therefore a breakthrough for clinicians who need to reliably detect and frequently monitor cognitive ability in MS patients to effectively treat sufferers. However, much more work must be done to refine the test for MS.

National screening

The smartphone app is indeed one way forward for a widespread national screening and remote monitoring programme for mild cognitive impairment, Alzheimer’s, and other forms of dementia in South Africa. It could certainly assist in identifying people with a high risk of developing the disease long before the appearance of symptoms so that appropriate steps can be taken to slow the progression of the disease.

It is apparent that in the Fourth Industrial Revolution smart technologies will increasingly transform the early detection and treatment of diseases.

The post Artificial intelligence detects dementia appeared first on Artificial Intelligence.

A new system uses artificial intelligence to detect errors in patients’ medication self-administration methods.

Artificial intelligence could help identify potential errors in a patient’s medication self-administration method, leading to reduced hospitalizations and healthcare costs, according to a study published in Nature Medicine.

Errors in medication self-administration lead to poor treatment adherence, increased hospitalizations, and higher care spending, researchers noted. These errors are especially common when medications involve devices like insulin pens or inhalers.

“Some past work reports that up to 70 percent of patients do not take their insulin as prescribed, and many patients do not use inhalers properly,” said Dina Katabi, the Andrew and Erna Viteri Professor at MIT.

Some common drugs also require intricate delivery mechanisms, making it difficult for patients to correctly administer medications themselves.

“For example, insulin pens require priming to make sure there are no air bubbles inside. And after injection, you have to hold for 10 seconds,” said Mingmin Zhao, a PhD student in MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL). “All those little steps are necessary to properly deliver the drug to its active site.”

Researchers developed a system that leverages artificial intelligence to reduce self-administration errors for some types of medications. The new tool uses wireless sensing and AI to determine when a patient is using an insulin pen or inhaler, and flags potential errors in the patient’s administration method.

The system works by using a sensor to track a patient’s movements within a ten-meter radius, using radio waves that reflect off their body. Then, AI analyzes the reflected signals for signs of a patient self-administering an inhaler or insulin pen. Finally, the system alerts the patient or their healthcare provider when it detects an error in the patient’s self-administration.

The team adapted their sensing method from a wireless technology they had previously used to monitor people’s sleeping positions. It starts with a wall-mounted device that emits low-power radio waves. When someone moves, they modulate the signal and reflect it back to the device’s sensor. Each unique movement yields a corresponding pattern of modulated radio waves the device can decode.

“One nice thing about this system is that it doesn’t require the patient to wear any sensors,” said Zhao. “It can even work through occlusions, similar to how you can access your Wi-Fi when you’re in a different room from your router.”

The new system can sit in the background at home, similar to a Wi-Fi router, and leverages AI to interpret the modulated radio waves. To train the AI algorithm, researchers performed example movements – some relevant, like using an inhaler, and some not, like eating. The system was able to detect 96 percent of insulin pen administration and 99 percent of inhaler uses.

After successfully detecting relevant movements, the system showed that it could detect errors as well. Because every proper medication administration follows a similar sequence, the system can flag anomalies in any particular step. For example, the system can recognize if a patient holds down their insulin pen for five seconds instead of the prescribed ten seconds. The system can then relay that information directly to the patient’s doctor so they can fix their technique.

“By breaking it down into these steps, we can not only see how frequently the patient is using their device, but also assess their administration technique to see how well they’re doing,” said Zhao.

A key feature of the system is its noninvasiveness, the team noted, which could encourage patients to actively participate in their own health.

“We think that the clinical implications of our system could be significant. We envision that this system will be able to provide continuous feedback for clinicians on their patients’ medication self-administration. Based on the feedback from our system, health professionals can then make a clinical judgment (for example, whether more training and education on medication device administration techniques is needed for the patient),” researchers stated.

“Additionally, this system could contribute to patient empowerment and engagement in their health by giving them feedback about their medication self-administration technique and allowing them to avoid common medication self-administration errors.”

The group also stated that the AI system could be adapted to medications beyond inhalers and insulin pens. Researchers would just have to re-train the algorithm to recognize the appropriate sequence of movements.

“With this type of sensing technology at home, we could detect issues early on, so the person can see a doctor before the problem is exacerbated,” Zhao concluded.

The post Artificial Intelligence Detects Medication Administration Errors appeared first on Artificial Intelligence.

Researchers have demonstrated that deep learning models can help neurologists interpret epileptic episodes during and between seizures from relatively few scalp electroencephalography (EEG) readings.

The best of the models proved out the concept under review: an automated annotation tool needing 142 times less EEG data than human experts would need to comb through—and epilepsy patients would need to log—using digital disease diaries.

The researchers, from IBM in collaboration with Temple University and other academic centers, worked with 87 scientists and software engineers from 14 research centers around the world to develop the models.

The research team analyzed EEG data from 365 patients representing 172,000 ictal (during a seizure) incidents and 2.2 million interictal (between seizures) occurrences. Part of the analysis was conducting a crowdsourced AI challenge.

“Participants were tasked to develop an ictal/interictal classifier with high sensitivity and low false alarm rates,” the study authors explain in a report published March 18 in the Lancet journal EBioMedicine. “We built a challenge platform that prevented participants from downloading or directly accessing the data while allowing crowdsourced model development.”

Led by IBM researchers Stefan Harrer in Australia and Gustavo Stolovitzky in the U.S., the team found their novel automated seizure detector returned sensitivities of up to 91.6%.

“This study demonstrates the utility of deep learning for patient-specific seizure detection in EEG data,” the authors comment. “Furthermore, deep learning in combination with a human reviewer can provide the basis for an assistive data labelling system lowering the time of manual review while maintaining human expert annotation performance.”

The post Deep learning detects, annotates epileptic seizures on scant EEG data appeared first on Artificial Intelligence.

Researchers may have developed a deep learning model that is effective at detecting electrolyte imbalance via electrocardiography (ECG).

“The detection and monitoring of electrolyte imbalance is essential for appropriate management of many metabolic diseases; however, there is no tool that detects such imbalances reliably and noninvasively,” the team wrote in their abstract.

Aiming to develop a deep learning model using ECG, the researchers conducted the retrospective cohort study of two hospitals. The patient sample included 92,140 patients who underwent a lab electrolyte exam and ECG within 30 minutes. The learning model was created using 83,449 ECGs of more than 48,000 of the patients (the internal validation cohort consisted of 12,091 ECGs from 12,091 patients), and they team conducted an external validation with the ECGs of more than 31,000 patients from another hospital. The researchers then evaluated the area under the receiving operating characteristic curve (AUC) of their deep learning model with the use of 12-lead ECG for detecting

According to the analysis results, the AUC for hyperkalemia was 0.945 and was 0.866 for hypokalemia. For hypernatremia, it was 0.944 and was 0.885 hyponatremia. For hypercalcemia, the AUC was 0.905, and for hypocalcemia, it was 0.901. Values during the external validation of the AUC for hypokalemia, hypernatremia, hyponatremia, hypercalcemia, and hypocalcemia were 0.873, 0.857, 0.839, 0.856, 0.831, and 0.813 respectively. The authors also reported that the learning model helped visualize the important ECG region for the detection of electrolyte imbalances.

“To the best of our knowledge, this study is the first to develop an artificial intelligence algorithm for detecting electrolyte imbalance and to show the interpretable patterns of decision making using artificial intelligence in the biosignal domain,” the authors wrote.

Some of the study limitations included the use of 4 common electrolytes (to the exclusion of others), the use of retrospective data, the limited number of centers, a lack of adjustment for certain comorbidities, and the limited combinations of ECG leads.

The post Deep Learning Model Detects Electrolyte Imbalance via ECG appeared first on Artificial Intelligence.

Machine learning tools were able to identify biomarkers in blood that could enable earlier diagnosis of children with autism spectrum disorder.

Machine learning tools analyzed hundreds of proteins and identified blood biomarkers that could speed the diagnosis of autism spectrum disorder (ASD), according to a study published in PLOS One.

ASD impacts at least one out of every 59 children in the US, researchers noted. The condition is also associated with significant personal, family, and societal costs. Efforts to determine the underlying biology of ASD, as well as ASD prevention, early diagnosis, and effective treatment, are public health priorities.

Being able to identify children with autism when they’re toddlers could make a big difference, the team stated. Currently, the average age of a child diagnosed with ASD in the US is four years old.

Dig Deeper

• Machine Learning, Facebook Data Offer Insight into Schizophrenia
• Machine Learning Tool Can Detect Changes in Serotonin Levels
• Machine Learning Scans Retinal Images to Predict Alzheimer’s Disease

Diagnosis before the age of four means that a child is more likely to receive an effective, evidence-based treatment, including therapies directed at core ASD symptoms like inflexible behaviors and lack of communication skills.

Researchers have investigated many blood-based biomarker candidates, including neurotransmitters, cytokines, and markers of mitochondrial dysfunction, oxidative stress, and impaired methylation. However, because ASD is so prevalent, using machine learning to incorporate demographic and clinical data into the analysis could more powerfully examine disease status and symptom severity.

For the study, the team examined serum samples from 76 boys with ASD and 78 from typically developing boys aged between 18 months and eight years. The results showed that all nine proteins in the biomarker panel were significantly different in boys with ASD compared with typically developing boys. Researchers found that each of the nine serum proteins correlated with symptom severity.

Researchers evaluated more than 1,100 proteins using the SomaLogic protein analysis platform. The group identified a panel of nine proteins as optimal for predicting ASD using three computational methods. Researchers then evaluated the biomarker panel for quality using machine learning methods.

“The more significantly affected the child is, the higher or lower than normal the blood biomarker is,” said Dwight German, PhD, professor of psychiatry at UT Southwestern and senior author of the study.

“Ideally, there will be a day when a child is identified using blood biomarkers as being at risk for developing ASD and therapies can be started immediately. That would help the child develop skills to optimize their communication and learning.”

The team noted that future studies will need to fully validate the present findings.

“Although the sample size is acceptable for a discovery study, the data presented here are preliminary, and a larger validation study is needed to be certain of the value of the biomarker panel. Due to the increased prevalence of ASD in boys, this study only enrolled boys, which does not allow for an investigation of gender-specific differences,” researchers noted.

The researchers expect that the study will pave the way for earlier diagnosis of autism.

“The earlier we can identify children with autism, the more understanding we can gain on ways to provide support and therapies that will improve their quality of life,” said Laura Hewitson, PhD, at The Johnson Center for Child Health & Development, a multidisciplinary treatment center in Austin, Texas, that uses a unique combination of clinical care, research, and education to further the understanding of ASD and related developmental disorders.

Researchers have previously turned to AI and data analytics techniques to enable earlier autism diagnosis. A study recently published in Nature Medicine showed that a precision medicine method enabled by artificial intelligence could lead to the first biomedical screening tool for a subtype of autism.

“Our study is the first precision medicine approach to overlay an array of research and healthcare data—including genetic mutation data, sexually different gene expression patterns, animal model data, EHR data, and health insurance claims data—and then use an AI-enhanced precision medicine approach to attempt to define one of the world’s most complex inheritable disorders,” said Yuan Luo, associate professor of preventive medicine: health and biomedical informatics at the Northwestern University Feinberg School of Medicine.

The post Machine Learning Detects Biomarkers of Autism Spectrum Disorder appeared first on Artificial Intelligence.