Analyzing sweat; automatic code repair; graphene in the brain.
Wearable sensors reveal health data from sweat
In the name of science, UC Berkeley researchers want you to break out into a sweat — so it can be analyzed, of course.
Specifically, the researchers have created a flexible sensor system that can measure metabolites and electrolytes in sweat, calibrate the data based upon skin temperature and sync the results in real time to a smartphone. They maintain this device is the first fully integrated electronic system that can provide continuous, non-invasive monitoring of multiple biochemicals in sweat in order to alert users to health problems such as fatigue, dehydration and dangerously high body temperatures.
Sweat is complex, and it is necessary to measure multiple targets to extract meaningful information about the state of a person’s health. As such, this is a fully integrated system that simultaneously and selectively measures multiple sweat analytes, and wirelessly transmits the processed data to a smartphone.
The prototype device contains five sensors on a flexible circuit board, which measure the metabolites glucose and lactate, the electrolytes sodium and potassium, and skin temperature. The integrated system allows for the use of the measured skin temperature to calibrate and adjust the readings of other sensors in real time, which is important because the response of glucose and lactate sensors can be greatly influenced by temperature.
Adjacent to the sensor array is the wireless printed circuit board with off-the-shelf silicon components. The researchers used more than 10 ICs to take the measurements from the sensors, amplify the signals, adjust for temperature changes and wirelessly transmit the data. To accompany this, the researchers developed an app to sync the data from the sensors to mobile phones, and fitted the device onto “smart” wristbands and headbands.
While this device works well on sweating athletes, there are likely to be many other applications of the technology for measuring vital metabolite and electrolyte levels of healthy persons in daily life including adaptation in order to monitor other body fluids for those suffering from illness and injury.
Correct code, automatically
In order to produce new repairs for a different set of programs, MIT researchers have developed a machine-learning system that can comb through repairs to open-source computer programs and learn their general properties.
The team tested the system on a set of programming errors, culled from real open-source applications, that had been compiled to evaluate automatic bug-repair systems. Where those earlier systems were able to repair one or two of the bugs, the MIT system repaired between 15 and 18, depending on whether it settled on the first solution it found or was allowed to run longer, they said.
While an automatic bug-repair tool would be useful in its own right, the work could have broader ramifications including identifying correct code.
Indeed, the researchers found that there are universal properties of correct code that can be learn from one set of applications and applied to another set of applications. If correct code can be recognized, that has enormous implications across all software engineering.
Graphene safely interacts with neurons in the brain
Using a material typically associated with electronics of the future, University of Trieste in Italy and the University of Cambridge researchers have shown that graphene can be used to make electrodes that can be implanted in the brain that may be able to be used to restore sensory functions for amputee or paralyzed patients, or for individuals with motor disorders such as Parkinson’s disease.
The researchers successfully demonstrated how it is possible to interface graphene – a 2D form of carbon – with neurons, or nerve cells, while maintaining the integrity of these vital cells. The work may be used to build graphene-based electrodes that can safely be implanted in the brain, offering promise for the restoration of sensory functions for amputee or paralyzed patients, or for individuals with motor disorders such as epilepsy or Parkinson’s disease.
Other groups have previously shown it is possible to use treated graphene to interact with neurons but the signal to noise ratio from this interface was very low. This team developed methods of working with untreated graphene, thereby retaining the material’s electrical conductivity, and making it a significantly better electrode.