Research news and notes

1. Hair helps repair nerves 

Repair of injury to the peripheral nerves has not changed much in many years. Direct suture of the severed nerve is only possible with clean acute cuts, so most cases require bridging of a gap. The criterion standard remains a nerve autograft. This is a rare commodity and can involve a separate incision to harvest the nerve with additional risks and complications. A biomaterial that can be as effective as an autograft is a long-sought goal in peripheral nerve surgery. Short gaps of a couple of centimeters have been bridged routinely with a variety of conduits, but there has been no breakthrough. In the January issue of Biomaterials, Seirpinski et al describe the use of a hydrogel made of keratin as a filler for a bridging conduit in nerve injury [4]. The keratin gel–filled graft resulted in a robust growth of axons across the gap, and activated Schwann cells. The motor outcome was not different from nerve autograft (it was actually poor in both cases at 6 weeks). In vitro, the keratin gel stimulated macrophages, which readily attached to keratin. So far, few biomaterials demonstrated neuroinductive properties. This is a preliminary study because the gap was only 4 mm long. However, if nerve growth can be demonstrated over longer gaps, this may be a paradigm-shifting discovery.

2. Deep brain stimulation may accelerate decisions and increase impulsivity 

Why is it often difficult to make a choice among apparently good options, such as making a selection from a restaurant menu? What is the neurologic basis of impulsivity? A recent report by Frank et al [1] provides interesting evidence about the importance of the subthalamic nucleus (STN) in delaying choices. Subthalamic nucleus appears to make it possible to delay decision until information in favor of one option or another crossed a threshold. The threshold varies according to the situation and the ability of the STN to function. In rat experiments, the animals that had STN dysfunction responded prematurely in choice tests. In this study, the authors tested patients with Parkinson's disease who had STN deep brain stimulation (DBS). The patients were tested with the stimulation either on or off. The computerized test presented letter combinations with hidden rules regarding which letters should be avoided and which should be chosen to score a “correct” choice. The rules were probabilistic, so uncertainty remained. The patients were first tested for memory and ability to learn the rules. For the conflict test that followed, the subjects were presented with letter combinations that were either high or low conflict, depending on the degree of certainty that the choice was correct based on prior experience.

Normally, subjects take longer to decide in the high-conflict situations. However, patients with DBS on did not exhibit this phenomenon—they were quicker than controls with their decisions in the high-conflict situation. In another part of the study, the researchers tested patients with Parkinson's disease who were treated with medication, either on or off the drugs. Dopaminergic medication impaired the patients' ability to learn from negative feedback. This finding tends to support the idea that dopamine mediates positive feedback and demonstrates that, in patients, it can disrupt negative feedback learning.

To better understand the neural networks that are involved in decision making and choices, the authors constructed a computer network that simulated the brain arrangement. They had an artificial cortex, striatum, thalamus, globus pallidus, substantia nigra, and STN. The model was able to accurately predict patient responses in each of the trials. One of the interesting phenomena the authors were able to test was the mechanism of action of DBS—does stimulation paradoxically act like a lesion, or does the stimulation actually enhance STN output? In the computer model, the authors either removed the “STN” unit or had it fire like a stimulator. Either way, the result was the disruption of the function of the unit in the network. The stimulation firing was not coordinated with the rest of the network and could not respond naturally and adaptively to outside signals.

This article represents a step in the ongoing revolution of our understanding of the workings of the brain. In earlier times, brain lesions taught us about the broad function of the different areas of the brain. More recent developments, such as DBS, functional magnetic resonance imaging, and advanced computer simulations, allow us to delve deeper into brain function. Learning more about the brain as an organ, we may come closer to some uncomfortable questions about the existence of free choice and the ability to manipulate or predict human behavior. On a different note, a neural network model that actually replicated the structure and connections of the human brain may be a fertile approach to artificial human-like intelligence. It makes sense that if you want to create a machine that imitates human thought, it should be built using the same organizational rules.

3. Genetic differences in the dopamine system influence the ability to learn from errors 

New discoveries are more trustworthy if they fit an emerging pattern. Just 1 month after the study of the cognitive effect of DBS and dopaminergic drugs, another study was published in Science looking at the same question from a completely different perspective [2]. Subjects with an allele for dopamine D2 receptor (A-1 allele) that results in a decreased density of these receptors were compared with noncarriers of that allele. They demonstrated lesser efficiency in learning to avoid actions with negative consequences. Functional magnetic resonance imaging demonstrated that their posterior medial frontal cortex responded less to negative feedback than in other subjects. Dynamic interactions between the medial frontal cortex and the hippocampus, thought to underlie feedback-based learning, were also decreased in A-1 allele carriers.

These findings are a startling demonstration of the importance of genetic predisposition in our behavior. A-1 allele carriers are known to have an increased risk of developing addictive behaviors. It is also interesting that dopamine, when given to patients with Parkinson's disease, decreased their ability to learn from negative feedback. This is a puzzling finding—maybe the administration of external dopamine down-regulates D2 receptors, rather than stimulate their function? We are just at the beginning of a new era of understanding of the biochemical and genetic basis of thought and behavior. These discoveries will affect us all.

4. Using behavioral economics to change patient behavior 

It is not entirely a coincidence that, in this era of increasing study of brain and behavior, a field like behavioral economics has developed. Last November, an interesting commentary was published in JAMA [3]. The authors addressed the perennial difficulty in getting patients to adopt behaviors that are beneficial to their health and to avoid those that are detrimental. They mentioned the failure of classic economics over account for the persistent irrational choices. Humans too often prefer immediate tangible benefits (a big tasty meal) to later less tangible benefits (avoidance of weight gain and its health hazards). Based on this recognition, the authors propose an approach that in behavioral economics is called asymmetric paternalism. As opposed to classic complete paternalism, asymmetric paternalism is intended to help those individuals who are prone to irrational decisions, while not limiting those who are making informed, deliberate decisions. Advocates of this approach use the same biases that prevent patients from making beneficial choices to steer them toward better options instead. For example, at fast-food restaurants, a bottle of water can be the default option for a combination meal, with a soft drink substituted upon request. Or, at a cafeteria, the more healthy choices can be placed first. Or the next routine checkup mammogram or colonoscopy can be automatically scheduled, with planned reminders, rather than leaving the scheduling to patients. The idea is to change the path of least resistance, or change the immediate effort or benefit related to healthy or unhealthy choices. This policy does not harm the freedom of choice because all options are preserved.

As our knowledge and understanding of human behavior increase, manipulation of human responses will become ever easier, while preserving the appearance of free choice. Who will decide what choices are better and therefore will be the default options? On the other hand, have we ever had a true freedom of choice?