Achanta et al. – iScience

To help tackle those issues, the team mapped the locations of those neurons in a rat’s heart in vivid 3D detail. “In effect what we have created is the first comprehensive roadmap of the heart’s nervous system that can be referenced by other researchers for a range of questions about the function, physiology, and connectivity of different neurons in the ICN,” said co-senior author Raj Vadigepalli, PhD. “The only other organ for which such a detailed high-resolution 3D map exists is the brain.”

Creating the model required two parallel approaches. In one, the team used a system called knife-edge scanning microscopy to create 750,000 images that could be assembled into a single 3D atlas. The second pipeline used “laser capture microdissection” to map individual neurons and sample their gene expression. Those were then placed in context with the map created from the first approach. “Because this hasn’t been done before, we were trouble-shooting the protocol as we went along,” said co-author Sirisha Achanta.

The 3D map showed that ICN neurons are located in clusters at the top of the heart where the veins and arteries attach. However, it confirmed that they can also be found in other places close to the sinoatrial node. “We know the sinoatrial atrial node is important in creating the heart rate or pace,” said co-author Dr. Jonathan Gorky. “Seeing the clustering of neurons around it was something we had always suspected but had never known for sure.”

The team also found sex-specific differences in neuron structures, a finding that could help explain the differences in heart disease between men and women. The team next plans to map a pig’s heart, which is closer in anatomy to a human.

However, the research is already paying dividends, as it is being used by other researchers learning how to improve the function of the heart and other organs. “Our protocol uses everyday lab materials and techniques,” said Achanta. “It is highly reproducible and is available now for other organ systems to map not just neurons, but other micro-structures.”

Typically, companies use GPUs as acceleration hardware in implementing deep learning in technology. But this is pricey — top of the line GPU platforms cost around $100,000. Rice researchers have now created a cost-saving alternative, an algorithm called sub-linear deep learning engine (SLIDE) that is able to do the same job of implementing deep learning, but without the specialized acceleration hardware.

The team then took a complex workload and fed it to both a top-line GPU using Google’s TensorFlow software, and a “44-core Xeon-class CPU” using SLIDE, and found the CPU could complete the training in just one hour, compared to three and a half hours for the GPU. (There is, to our knowledge, no such thing as a 44-core Xeon-class CPU, so it’s likely that the team is referring to a 22-core, 44-thread CPU.)

SLIDE works by taking a fundamentally different approach to deep learning. GPUs leverage such networks by studying huge amounts of data — often using millions or billions of neurons, and employing different neurons to recognize different types of information. But you don’t need to train every neuron on every case. SLIDE only picks the neurons that are relevant to the learning at hand.

According to Anshumali Shrivastava, assistant professor at Rice’s Brown School of Engineering, SLIDE also has the advantage of being data parallel. “By data parallel I mean that if I have two data instances that I want to train on, let’s say one is an image of a cat and the other of a bus, they will likely activate different neurons, and SLIDE can update, or train on these two independently,” he said. “This is much a better utilization of parallelism for CPUs.”

This did bring its own challenges, however. “The flipside, compared to GPU, is that we require a big memory,” he said. “There is a cache hierarchy in main memory, and if you’re not careful with it you can run into a problem called cache thrashing, where you get a lot of cache misses.” After the team published their initial findings, however, Intel got in touch to collaborate on the problem. “They told us they could work with us to make it train even faster, and they were right. Our results improved by about 50 percent with their help.”

SLIDE is a promising development for those involved in AI. It’s unlikely to replace GPU-based training any time soon, because it’s far easier to add multiple GPUs to one system than multiple CPUs. (The aforementioned $100,000 GPU system, for example, has eight V100s.) What SLIDE does have, though, is the potential to make AI training more accessible and more efficient.

Shrivastava says there’s much more to explore. “We’ve just scratched the surface,” he said. “There’s a lot we can still do to optimize. We have not used vectorization, for example, or built-in accelerators in the CPU, like Intel Deep Learning Boost. There are a lot of other tricks we could still use to make this even faster.” However, the key takeaway, Shrivastava says, is that SLIDE shows there are other ways to implement deep learning. “Ours may be the first algorithmic approach to beat GPU, but I hope it’s not the last.”

Scientists already knew that humans have specific areas of the brain used to recognize faces, words, numbers and even celebrities like Jennifer Aniston, Bill Clinton and Kobe Bryant. Those areas of the brain are often in the same place for large groups of people. A recent study by Harvard Medical School found that for monkeys to develop similar regions, they need to be exposed to objects from a young age. The researchers at Stanford wondered if the same were true in humans. Pokémon fans proved to be perfect subjects, as most began playing when they were quite young and because the characters are so unique.

The findings support theories that early childhood exposure is necessary for developing dedicated brain regions, and that from an early age, our brains change in response to experiential learning. Because the fold activated by characters like Pikachu, Wobbuffet and Bulbasaur is the same fold that responds to images of animals, the researchers also believe there’s some kind of underlying constraints hardwired into the brain that determine how those changes happen.

Now, as you watch the long-awaited Detective Pickachu movie coming out later this week, you might wonder if you have a dedicated brain fold actively identifying the characters. You might also wonder how other games you obsessed with as a kid have found their way into the structure of your brain.