Neural networks can create fake fingerprints to fool biometric scanners, while elsewhere researchers are printing electric circuits onto temporary tattoos. This week, all the coolest discoveries fit to print also include a hopeful Parkinson’s treatment, better detection of food contamination, and a data-collection project that’s diving deep into the bowels of the world’s … bowels.
What is it? In patients with Parkinson’s disease, a loss of dopamine-producing neurons in the brain leads to symptoms like bodily tremors, stiff muscles and difficulty walking. Now neurosurgeons in Japan have implanted “reprogrammed” cells into the brain of a Parkinson’s patient that might make up for the neuron shortage and alleviate symptoms.
Why does it matter? This is the first time this kind of therapy, which relies on what are called induced pluripotent stem cells, has been tried on a Parkinson’s patient.
How does it work? Scientists developed induced pluripotent stem cells, or iPS cells, by taking regular adult cells, such as skin or blood cells, and “reprogramming” them into a state where they resemble stem cells — undifferentiated cells that can develop into many different cell types in the body. Working out of Kyoto University Hospital, scientists used the technique “to transform iPS cells into precursors to the neurons that produce the neurotransmitter dopamine,” according to a story in the journal Nature describing the therapy. In October, neurosurgeon Takayuki Kikuchi implanted 2.4 million of the dopamine precursor cells into the brain of a Parkinson’s patient in his 50s, aiming his implants specifically at 12 sites known to be centers of dopamine activity. The team plans to test the technique on six more patients by 2020; if the therapy proves effective, it could become commercially available as soon as 2023.
What is it? A partnership between scientists at Carnegie Mellon University and Portugal’s University of Coimbra has yielded an ultrathin electric circuit that can be applied to the skin just like a temporary tattoo — because it’s printed on the same kind of paper as temporary tattoos.
Why does it matter? This new tech was developed as part of a larger collaboration between the two universities. The project, called Stretchtronics, is looking at ways to embed wearable low-cost electronics onto thin, flexible materials that can be attached to the skin, where they could monitor things like heart rate and muscular activity.
How does it work? Researchers use an inkjet printer to print circuits of silver nanoparticles onto the tattoo paper. They coat those circuits with a metal alloy that fuses with the silver and both increases conductivity and helps the circuit stretch — in fact, the stretchability of these tattoo circuits is similar to that of human skin. “This is a breakthrough in the printed electronics area,” said Coimbra’s Mahmoud Tavakoli, a researcher on the project. The tattoos can be applied with water. The team reported its findings in Advanced Materials.
What is it? Perhaps you’ve heard of the Svalbard Global Seed Vault, where scientists wanting to protect plant biodiversity and plan against future disaster keep a huge bank of seeds snuggled into the Norwegian permafrost. Another set of scientists — who probably wash their hands more often — are doing a similar kind of thing, but with human feces.
Why does it matter? The Global Microbiome Conservancy was founded in 2016 by several MIT microbiologists who observed that, as traditional societies around the world increasingly adapt their diets to Western influence (processed foods, antibiotics), we’re seeing in a decline in the diversity of human microbiomes — the rich worlds of bacteria in our guts. That decline, in turn, presents a health threat; a few studies have found an absence of certain diseases in societies with diverse gut microbiota. The microbiologists set out to rectify this loss by gathering a diverse worldwide collection of human microbiomes, best accomplished via fecal sampling. As GMC co-founder Eric Alm told Science magazine, the microbes could help scientists find treatments for both gut diseases and other health problems like asthma and obesity: “I’m 100% confident that there are relevant medical applications for hundreds of strains we’ve screened and characterized.”
How does it work? The, uh, old-fashioned way. Researchers travel the world asking people to do their business in an assigned bowl, whence the samples are divided into groups: some to be dried and DNA-sequenced, others to be shipped frozen back to Cambridge, where discrete bacterial strains can be isolated. Already MIT scientists have uncovered 55 previously unknown genera of bacteria from 4,000 strains collected from Africa and the Arctic. On the other hand, genetic collection of this sort may pose ethical challenges, which the Science article explores.
What is it? New York University researchers announced that they’d used a neural network to generate artificial fingerprints that could trick biometric identification systems. Described as a “master key” for fingerprint-ID systems, these “DeepMasterPrints” could fool such systems about one in five times, versus a typical error rate of one in a thousand.
Why does it matter? Besides being a plot point perfectly suited to the next “Mission Impossible” movie, the finding should actually help security experts, in the way that hackers can help cybersecurity squads — by identifying possible holes that need to be patched up. The NYU researchers presented a paper on their findings (PDF) at a security conference in Los Angeles.
How does it work? By taking advantage of two properties common to fingerprint-ID systems, according to the Guardian: the fact that scanners typically read only partial, as opposed to whole, fingerprints, comparing the partial scan with partial existing records; and the fact that “some features of fingerprints are more common than others.” Exploiting these weaknesses, the neural network used a common technique called a “generative adversarial network” to create fake fingerprints that looked enough like real ones; as the Guardian notes, the technique wouldn’t work to target a specific account, but if it were used at scale, hackers might be able to access enough accounts to make it worth their while.
What is it? Radio-frequency identification tags are a ubiquitous technology used to track billions of consumer products. Now scientists at MIT think they could use RFID tags — with no hardware modification required — to detect signs of contamination in food products.
Why does it matter? If used broadly, the technology could put the power to detect food contamination into the hands of consumers, and thereby reduce foodborne illness spread by contaminated products. An MIT press release points to the 2008 hospitalization, in China, of 50,000 babies who’d consumed formula tainted with melamine, and the more recent deaths of 100 people in Indonesia from drinking alcohol spiked with methanol, a toxic substance. In their study, presented at the ACM Workshop on Hot Topics in Networks, the researchers tested the technology’s efficacy on baby formula and alcohol.
How does it work? When attached to consumer products, RFID tags power up and emit electromagnetic waves when they’re pinged by a wireless reader. The waves they emit, though, “travel into and are distorted by the molecules and ions of the contents in the container,” according to MIT — so changes in the contents will be reflected in the emitted waves. For instance, if a container is empty, the RFID will respond at around 950 megahertz, but if it’s filled with water, which absorbs some of the frequency, the tag’s response will be around 720 megahertz. Hoping to exploit this kind of differential, MIT researchers sought to create a machine-learning model that could flag potential contamination in products — which it did in melamine-laced baby formula with 96 percent accuracy, and with methanol-laced alcohol with 97 accuracy.