Let sleeping dogs learn
Humans show particular thalamo-cortical activity and the associated EEG-pattern during sleep. These so called sleep spindles are associated with integration of new knowledge and management as well as consolidation of memories. The goal of a recent study by Iotchev et al. (2017) was to determine whether dogs, man’s best friend, show similar patterns and activity. And indeed, Iotchev et al. found that the density of sleep spindles were linked to memory and subsequently led to better performance in a recall session. We can conclude that dogs show functional equivalents to human sleep spindles and therefore say that dogs solidify memories in their sleep just like us. So there is some reason behind the expression “let sleeping dogs lie”.
Specialized macrophages promote obesity
Obesity is a medical condition in which excess body fat accumulates and therefore has a negative impact on health. Many experts view it as one of the most serious public health problems of the 21st century and it is one of the leading preventable causes of death worldwide. In a new study by Pirzgalska et al. (2017), crucial progress has been made in identifying the underlying mechanisms of obesity.
Macrophages, white blood cells that dispose of foreign substances in body tissue, have long been identified as mediators of tissue inflammation in obesity. Pirzgalska et al. were now able to show that especially sympathetic neuron-associated macrophages (SAMs) play a crucial role in noradrenaline-regulated thermogenesis, a process altered in obesity. SAMs are able to import and degrade noradrenaline and have a higher noradrenaline level as compared to adipose tissue macrophages (ATMs). The study then proceeds to show that the level of SAMs was significantly higher in obese mouse models than in thin mice. All in all, the findings of Pirzgalska et al. suggest that SAMs facilitate obesity through noradrenaline clearance and these macrophages could be potentially targeted in further studies looking for a therapy for obesity.
Artificial intelligence teaches itself
Researchers from DeepMind in London have been able to reach a main goal of artificial intelligence. This goal is to create an algorithm that learns, starting from scratch, extraordinary proficiency, like it was in the case of AlphaGo. This AI was the first computer program to defeat a world champion in the abstract and complex board game GO. Now, this AI has been modified with a new algorithm entirely based on reinforcement learning, which lacks any input of human guidance or knowledge. The modified AI, called AlphaGo Zero, teaches itself and has been able to beat the champion-vanquishing AI AlphaGo 100 to 0. Over thousands of years and millions of games, humanity has congregated knowledge about GO and AlphaGo zero was able to accumulate much of this knowledge as well as new strategies in just a few days.
More about artificial intelligence in the next issue of The Cambridge Student and online on tcs.cam.ac.uk
Iotchev, I.B., Kis, A., Bodizs, R., van Luijtelaar, G., & Kubinyi, E. (2017) EEG Transients in the Sigma Range during non-REM sleep predict learning in dogs. Scientific Reports 7: 12936 | DOI:10.1038/s41598-017-13278-3
Silver, D., Schrittwieser, J., Simonyan, K., Antonoglou, I., Huang, A., Guez, A., Hubert, T., Baker, L., Lai, M., Bolton, A., Chen, Y., Lillicrap, T., Hui, F., Sifre, L., van den Driessche, G., Graepel, T., & Hassabis, D. (2017) Mastering the game of Go without human knowledge. Science. Vol 550. P. 354 – 358. doi:10.1038/nature24270
Pirzgalska, R.M., Seixas, E., Seidman, J.S., Link, V.M., Sanchez, N.M., Mahu, I., Mendes, R., Gres, V., Kubasova, N., Morris, I., Arus, B.A., Larabee, C.M., Vasques, M., Tortosa, F., Sousa, A.L., Anandan, S., Tranfield, E., Hahn, M.K., Iannacone, M., Spann, N.J., Glass, C.K., & Domingos, A.I. (2017) Sympathetic neuron-associated macrophages contribute to obesity by importing and metabolizing norepinephrine. Nature Medicine. doi: 10.1038/nm.4422