This Special Music Helped Preemie Babies’ Brains Develop
Listening to music helped preterm babies' brains develop, according to the results of a new Swiss study.
Move over, Baby Einstein: New research from Switzerland shows that listening to soothing music in the first weeks of life helps encourage brain development in preterm babies.
For the study, the scientists recruited a harpist and a new-age musician to compose three pieces of music.
The Lowdown
Children who are born prematurely, between 24 and 32 weeks of pregnancy, are far more likely to survive today than they used to be—but because their brains are less developed at birth, they're still at high risk for learning difficulties and emotional disorders later in life.
Researchers in Geneva thought that the unfamiliar and stressful noises in neonatal intensive care units might be partially responsible. After all, a hospital ward filled with alarms, other infants crying, and adults bustling in and out is far more disruptive than the quiet in-utero environment the babies are used to. They decided to test whether listening to pleasant music could have a positive, counterbalancing effect on the babies' brain development.
Led by Dr. Petra Hüppi at the University of Geneva, the scientists recruited Swiss harpist and new-age musician Andreas Vollenweider (who has collaborated with the likes of Carly Simon, Bryan Adams, and Bobby McFerrin). Vollenweider developed three pieces of music specifically for the NICU babies, which were played for them five times per week. Each track was used for specific purposes: To help the baby wake up; to stimulate a baby who was already awake; and to help the baby fall back asleep.
When they reached an age equivalent to a full-term baby, the infants underwent an MRI. The researchers focused on connections within the salience network, which determines how relevant information is, and then processes and acts on it—crucial components of healthy social behavior and emotional regulation. The neural networks of preemies who had listened to Vollenweider's pieces were stronger than preterm babies who had not received the intervention, and were instead much more similar to full-term babies.
Next Up
The first infants in the study are now 6 years old—the age when cognitive problems usually become diagnosable. Researchers plan to follow up with more cognitive and socio-emotional assessments, to determine whether the effects of the music intervention have lasted.
The first infants in the study are now 6 years old—the age when cognitive problems usually become diagnosable.
The scientists note in their paper that, while they saw strong results in the babies' primary auditory cortex and thalamus connections—suggesting that they had developed an ability to recognize and respond to familiar music—there was less reaction in the regions responsible for socioemotional processing. They hypothesize that more time spent listening to music during a NICU stay could improve those connections as well; but another study would be needed to know for sure.
Open Questions
Because this initial study had a fairly small sample size (only 20 preterm infants underwent the musical intervention, with another 19 studied as a control group), and they all listened to the same music for the same amount of time, it's still undetermined whether variations in the type and frequency of music would make a difference. Are Vollenweider's harps, bells, and punji the runaway favorite, or would other styles of music help, too? (Would "Baby Shark" help … or hurt?) There's also a chance that other types of repetitive sounds, like parents speaking or singing to their children, might have similar effects.
But the biggest question is still the one that the scientists plan to tackle next: Whether the intervention lasts as the children grow up. If it does, that's great news for any family with a preemie — and for the baby-sized headphone industry.
Artificial Intelligence Needs Doctors As Much As They Need It
In this futuristic medical concept, a doctor assesses a patient with robust machine assistance.
The media loves to hype concerns about artificial intelligence: What if machines become super-intelligent and self-aware? How will humanity compete and survive? But artificial intelligence today is a far cry from a robot takeover. "AI" is a catch-all term that often refers to machine training or machine learning: There is an abundance of data, vastly more than the human mind can assimilate, being tagged, captured and stored. This systematic data processing requires methodologies that can put it in usable form and formats. While these new developments stoke fear in some corners, the ability to predict outcomes is generally seen as a good thing, as it can mitigate risks and even save lives.
We, collectively, want AI even though it is seldom expressed this way.
The prospects and attempts toward artificial intelligence has been with us for decades. Only recently have the underlying technologies and infrastructure--including computer processing, storage, networking speed and advanced software platforms--become omnipresent. These technological advances enabled the implementation of data mining concepts and the subsequent advantages that were not feasible just a decade ago.
AI is fantastical by vision, evolutionary by experience, and disruptive upon reflection. In the world of health care, AI is already transforming research and clinical practice. We, collectively, want AI even though it is seldom expressed this way. What we, the patient population, patient advocates and caregivers, agree on and want is: (1) timely, precise and inexpensive diagnoses of our ailments, injuries and disorders; (2) timely, personalized, highly effective and efficient courses of therapies; and (3) expedited recovery with minimum deficits, complications and recurrence.
"Artificial intelligence and machine learning will impact healthcare as profoundly as the discovery of the microscope."
Implicitly, we all are saying that we want our healthcare systems and clinicians to accomplish truly inhuman feats: to incorporate all sources of structured data (such as published statistics and reports) and unstructured data (including news articles, conversational analysis by care givers, nuances of similar cases, talks at professional societies); to analyze the data sourced and uncover patterns, reveal side effects, define probable success and outcomes; and to present the best personalized course of treatment for the patient that addresses the ailment and mitigates associated risks. It is hard to argue against any of this.
In a recent published interview, Keith J. Dreyer, executive director of the Massachusetts General Hospital and Brigham and Women's Hospital Center for Clinical Data Science, says that "artificial intelligence and machine learning will impact healthcare as profoundly as the discovery of the microscope."
But as AI helps physicians in profound ways, like detecting subtle lesions on scans or distinguishing the symptoms of a stroke from a brain tumor, we humans can't get too complacent. Evolving AI platforms will provide more sophisticated sets of "tools" to address both mundane and complex medical challenges, albeit with humans very much in the mix and routinely at the helm.
Humans do not appear endangered to be replaced anytime soon.
Human beings are capable of a level of nuance and contextual understanding of complex medical scenarios and, consequently, do not appear endangered to be replaced anytime soon. These platforms will do some heavy lifting for sure and provide considerable assistance across the healthcare industry. But human involvement is crucial, as we are best at adaptive learning, cognition, ensuring accuracy of the data, and continually providing feedback to improve the machine learning components of the AI platforms that the health industry will increasingly rely upon.
The human/machine interface is not binary; there is no line in the sand. It is fuzzy and evolutionary, a synchronicity that we all will surely witness and experience. In the future, it may be possible that all recorded knowledge, including genetic, genomic and laboratory data, from structured and unstructured sources, can be at the fingertips of your clinician, and then factored into diagnosing your condition and prescribing your course of treatment. This is precision and personalized medicine on a grand scale applied at the micro level--you!
But none of this will diminish the importance of doctors, nurses and all assortment of care providers. Though they all will undoubtedly become more effective with such awesome AI assistance, their job will always be to heal you with compassion, wisdom, and kindness, for the essence of humanity cannot be automated.
This Revolutionary Medical Breakthrough Is Not a Treatment or a Cure
A doctor examines a patient to determine the cause of her illness.
What is a disease? This seemingly abstract and theoretical question is actually among the most practical questions in all of biomedicine. How patients are diagnosed, treated, managed and excused from various social and moral obligations hinges on the answer that is given. So do issues of how research is done and health care paid for. The question is also becoming one of the most problematic issues that those in health care will face in the next decade.
"The revolution in our understanding of the human genome, molecular biology, and genetics is creating a huge--if little acknowledged--shift in the understanding of what a disease is."
That is because the current conception of disease is undergoing a revolutionary change, fueled by progress in genetics and molecular biology. The consequences of this shift in the definition of disease promise to be as impactful as any other advance in biomedicine has ever been, which is admittedly saying a lot for what is in essence a conceptual change rather than one based on an empirical scientific advance.
For a long time, disease was defined by patient reports of feeling sick. It was not until the twentieth century that a shift occurred away from subjective reports of clusters of symptoms to defining diseases in terms of physiological states. Doctors began to realize that not all symptoms of fever represented the presence of the same disease. Flu got distinguished from malaria. Diseases such as hypertension, osteoporosis, cancer, lipidemia, silent myocardial infarction, retinopathy, blood clots and many others were recognized as not producing any or slight symptoms until suddenly the patient had a stroke or died.
The ability to assess both biology and biochemistry and to predict the consequences of subclinical pathological processes caused a distinction to be made between illness—what a person experiences—and disease—an underlying pathological process with a predictable course. Some conditions, such as Gulf War Syndrome, PTSD, many mental illnesses and fibromyalgia, remain controversial because no underlying pathological process has been found that correlates with them—a landmark criterion for diagnosing disease throughout most of the last century.
"Diseases for which no relationship had ever been posited are being lumped together due to common biochemical causal pathways...that are amenable to the same curative intervention."
The revolution in our understanding of the human genome, molecular biology, and genetics is creating a huge--if little acknowledged--shift in the understanding of what a disease is. A better understanding of the genetic and molecular roots of pathophysiology is leading to the reclassification of many familiar diseases. The test of disease is now not the pathophysiology but the presence of a gene, set of genes or molecular pathway that causes pathophysiology. Just as fever was differentiated into a multitude of diseases in the last century, cancer, cognitive impairment, addiction and many other diseases are being broken or split into many subkinds. And other diseases for which no relationship had ever been posited are being lumped together due to common biochemical causal pathways or the presence of similar dangerous biochemical products that are amenable to the same curative intervention, no matter how disparate the patients' symptoms or organic pathologies might appear.
We used to differentiate ovarian and breast cancers. Now we are thinking of them as outcomes of the same mutations in certain genes in the BRCA regions. They may eventually lump together as BRCA disease.
Other diseases such as familial amyloid polyneuropathy (FAP) which causes polyneuropathy and autonomic dysfunction are being split apart into new types or kinds. The disease is the product of mutations in the transthyretin gene. It was thought to be an autosomal dominant disease with symptomatic onset between 20-40 years of age. However, as genetic testing has improved, it has become clear that FAP's traditional clinical presentation represents a relatively small portion of those with FAP. Many patients with mutations in transthyretin — even mutations commonly seen in traditional FAP patients — do not fit the common clinical presentation. As the mutations begin to be understood, some people that were previously thought to have other polyneuropathies, such as chronic inflammatory demyelinating neuropathy, are now being rediagnosed with newly discovered variants of FAP.
"We are at the start of a major conceptual shift in how we organize the world of disease, and for that matter, health promotion."
Genome-wide association studies are beginning to find many links between diseases not thought to have any connection or association. For example some forms of diabetes, rheumatoid arthritis and thyroid disease may be the products of a small family of genetic mutations.
So why is this shift toward a genetic and molecular diagnostics likely to shake up medicine? One obvious way is that research projects may propose to recruit subjects not according to current standards of disease but on the basis of common genetic mutations or similar errors in biochemical pathways. It won't matter in a future study if subjects in a trial have what today might be termed nicotine addiction or Parkinsonism. If the molecular pathways producing the pathology are the same, then both groups might well wind up in the same trial of a drug.
In addition, what today look like common maladies—pancreatic cancer, severe depression, or acne, for example, could wind up being subdivided into so many highly differentiated versions of these conditions that each must be treated as what we now classify as a rare or ultra-rare disease. Unique biochemical markers or genetic messages may see many diseases broken into a huge number of distinct individual disease entities.
Patients may find that common genetic pathways or multiple effects from a single gene may create new alliances for fund-raising and advocacy. Groups fighting to cure mental and physical illnesses may wind up forgetting about their outward differences in the effort to alter genes or attack common protein markers.
Disease classification appears stable to us—until it isn't. And we are at the start of a major conceptual shift in how we organize the world of disease, and for that matter, health promotion. Classic reductionism, the view that all observable biological phenomena can be explained in terms of underlying chemical and physical principles, may turn out not to be true. But the molecular and genetic revolutions churning through medicine are illustrating that reductionism is going to have an enormous influence on disease classification. That is not a bad thing, but it is something that is going to take a lot to get used to.