The Cerebellum Holds Many Clues for Creating Humanoid Robots

When you visualize someone moving like a robot, what images come to mind? A lack of fluidity and coordination marked by awkward and jerky movements are cliché characteristics of robotic movements. Interestingly, all of these traits are linked to diseases or deficits in the human cerebellum which is responsible for coordinating fine-tuned muscle movements, balance, posture and much, much more.

Stereotypically, a robot moves and speaks as if it suffers from ataxia, which is a lack of muscle control during voluntary movements—such as walking, grasping, and picking up objects. The cerebellum is directly linked to ataxia, which can also affect the fluidity of speech, eye movements, and swallowing.

The cerebrum, which is the house of the prefrontal cortex and the seat of executive function, holds explicit knowledge beautifully. The cerebellum—which is the house of Purkinje neurons and the seat of muscle memory—holds implicit know-how and things that we learn to do through trial-and-error, repetition, and practice, practice, practice.

I've been researching the cerebellum for over a decade. From my perspective, it seems obvious that the secret to creating "ataxia-free" robots lies in applying the functions of the cerebellum to advanced artificial intelligence (AI) robotics.

A robot with a fully-functioning cerebellum could move with fluidity, stay balanced, and be proprioceptive. It could also think, communicate socially, learn implicitly, and have sensorimotor independence. I believe that a humanoid robot with a "brain" that equally represented both the cerebrum and the cerebellum could become a sentient android.

A Brief History of the Cerebellum

Aristotle (384-322 BC) is the first known writer to describe the cerebellum as ”parencephalon” and the cerebrum as the encephalon. In 1504 Leonardo da Vinci made wax castings of the human brain and coined the term “cerebellum” which is Latin for “little brain.” The “cerebrum” or “big brain” sits atop the cerebellum which is neatly tucked under the cerebrum.

The cerebellum is only 10 percent of brain volume but holds over 50 percent of your brain’s total neurons. Based on this disproportionate distribution of neurons, my father, Richard Bergland—who was a neuroscientist, neurosurgeon and author of The Fabric of Mind—always said, “We don't know exactly what the cerebellum is doing, but whatever it's doing, it’s doing a lot of it.”

Throughout history, the cerebellum has remained mysterious and under the radar to most neuroscientists. In the late 1960s and early 1970s there was a surge of interest in the cerebellum led by people like David Marr, John Eccles, and James Albus (who were my father’s peers) ... but interest in the cerebellum waned during the later part of the 20th century.

Luckily, neuroscientific interest in the cerebellum is currently having a renaissance and gaining momentum. In recent months, there has been a watershed of revolutionary discoveries about the cerebellum. This is a very exciting time to be researching the cerebellum.

The Cerebellum Could Take Center Stage in the Near Future

Although I’ve had my antennae up for studies about the cerebellum for decades, the breakneck speed of recent findings has left my head spinning. At the 2014 Society of Neuroscience conference last November, the cerebellum seemed to be trending heavily. In the last year alone, four different studies were released linking the cerebellum with autism spectrum disorders (ASD).

Based on the recent link between the cerebellum and autism it seems the same symptoms of ASD, such as difficulty using nonverbal communication skills, making eye contact, interpreting facial expressions, and motor skill deficits, could be the result of problems in the cerebellum.

People with autism also often have trouble interpreting another person's feelings, or being compassionate to other's pain and suffering. Ironically, rudimentary robots have been used to help children with autism become more socially engaged, according to research from the Yale Child Study Center.

Another study earlier this year linked the cerebellum with bipolar disorders. And now the cerebellum is being linked to optimized brain-interfaced robotics. Could the key to creating a sentient robot with feelings be linked to creating an AI cerebellum structure? I suspect the answer is yes.

Until a few weeks ago, I hadn’t thought much about how a better understanding of the cerebellum could be applied to robotics or brain-interfaces that coordinate the movement of bionic prosthetics. That all changed recently when I read a February 2015 press release by Scott Frey at the University of Missouri about his study that was published in the Journal of Cognitive Neuroscience.

Scott Frey and I have spent a couple hours speaking on the phone about his research. It turns out we both are former triathletes who share a passion for the transformative power of physical activity, neuroscience, and the cerebellum.

In the press release about his discovery that the cerebellum plays a critical role in helping subjects grasp an object using a brain-controlled interface Scott Frey said,

"We live in a world of advanced technology in which a button can move a crane or open a door. For those with disabilities, assistive technologies, such as robotic arms or sensors inserted in the brain, make it possible to accomplish actions like grasping with the press of a button or directly through brain activity; however, little is known about how the human brain adapts to these technologies. We found that the brain didn't necessarily evolve to control modern robotic arms, but rather the cerebellum, an ancient portion of our brain that has remained relatively unchanged, plays a vital role in helping us reach and grasp with these tools--often with only minimal training."

Over the past few days I have unearthed multiple other studies that support the potential role that cerebellar devices could play in the creation of humanoid robots and androids.

Traditionally, computers and AI robotics focus primarily on the “cerebral” aspects of intelligence held in the cerebrum. (“Cerebellar” is the the sister word to cerebral and defines anything “relating to or associated with the cerebellum.”)

What makes a humanoid robot so much more complex than a computer like IBM's “Watson”—which is called a "cognitive computer" with cerebral intelligence and can win a game show like Jeopardy—is that a sentient robot needs to have the ability to move through space and have the sensorimotor abilitiy to adapt to an ever-changing environment. This ability would appear to require the special capabilities of a cerebellum.

During the 1970s, when Jamie Somers and Steve Austin battled the “fembots” on The Bionic Woman and Six Million Dollar Man, the biggest problem they faced was that the fembots—which were human-like androids—were stronger than either of their partial bionic prosthetics. The only real disadvantage the fembots had was that they couldn’t think for themselves.

Incorporating cerebellar models into the design of humanoid robots could be central to creating a sentient robot that can both feel emotions and think for itself.

Why Am I So Personally Invested in Advocating for the Powers of the Cerebellum?

In his work, my dad was obsessed with the cerebellum and the potential for “cerebellar genius” that he considered was a type of intelligence learned through practice and experience. My father was an avid tennis player. To illustrate the link between the cerebrum and the cerebellum he would say, "Of this I am absolutely positive; becoming a neurosurgeon was the direct consequence of my eye for the ball.”

The cerebellum is responsible for your Vestibulo-Ocular Reflex (VOR) which allows your eyes to track a target while your head moves in various directions. As a neurosurgeon my father had to apply both cerebral and cerebellar skills, as every surgeon, athlete, and performer must.

As I was growing up, the only time that my father and I ever really had quality time together was either on the tennis court or playing chess. From early on my father planted the seeds of a split-brain model that implied the importance of “cerebral” smarts that you could learn from books and the type of executive function one uses while sitting still and playing a strategic game of chess or taking a test. But he noted the importance of "cerebellar smarts" as well.

"Cerebellar smarts” are gained through trial-and-error, life experience and regular practice on the court, or in the field. The cerebellum is central to mastering every type of athletics. I have come to believe that cerebral and cerebellar intelligence are equally important and should always go hand-in-hand both in life and in sport.

Because of my father’s influence, I became a champion for the cerebellum and believe that it has traditionally been undervalued. I firmly believe that the cerebellum plays a much bigger role in cognitive function and “intelligence” than—until relatively recently—most of academia or neuroscience has seemed ready to admit.

Non Satis Scire: "To Know Is Not Enough"

I went to a small liberal arts college in Amherst, Massachusetts called Hampshire College that has no tests or grades and does not look at SAT scores during the admissions process. I am terrible at taking tests and had horrible SAT scores, which is the main reason I went to Hampshire. It was literally the only college that would accept me.

The motto of Hampshire is Non Satis Scire which is Latin for "To know is not enough." I believe that, when it comes to both education and building AI robots, that it's important to remember that—based on the ratio of neurons between the cerebrum and cerebellum, and common sense—cerebral smarts are less than half of our working human intelligence.

As the father of a 7-year-old, I am disturbed to see the emphasis and pressure schools put on children through the Common Core Standards and No Child Left Behind to do well on standardized tests while depriving them of the physical activity and creative outlets the cerebellum needs to stay connected to the cerebrum and to thrive.

When I was writing The Athlete's Way (St. Martin's Press) a decade ago, I created a split-brain model called "up brain-down brain" that challenged the conventional ideas of "left brain-right brain." In my humble opinion, the salient divide in the cranial globe is north-south, between the cerebrum and the cerebellum. If you'd like to read more about this split-brain model please click here for a free sample from my book.

From athletics, to child development, or rehabilitation from a stroke, putting the emphasis on the left-right and up-down harmony and interconnection of both brain hemispheres of the cerebrum and both hemispheres of the cerebellum is the key to success and is the foundation of the theories presented in The Athlete’s Way.

Logically, if people of all ages create daily habits that optimize the white matter connectivity, and gray matter volume, of both hemispheres of the cerebrum (up brain) and cerebellum (down brain) they can improve their odds of achieving a lifespan of personal bests. This is a simple and road-tested prescriptive which is the foundation of my current book-in-progress titled Superfluidity.

The cerebellum is an underdog, and a mighty mouse, that ounce for ounce packs a walloping punch. Cerebellar genius may ultimately be more important than cerebral intelligence in terms of our human species' survival and needs to be considered when trying to build sentient androids.

For example, in those instances when there is an “unexpected error” and you are caught off-guard by a slip on the ice, or a near miss while driving your car, it is your cerebellum that instantaneously takes over and predicts and coordinates each of the complex muscle movements that keep you unharmed and alive, before your cerebrum even has time to think.

Recently, researchers at Penn identified how this works in a fascinating study about climbing fibers, granule, and Purkinje cells all working together with lightning fast precision. The cerebellum has the exhausting job of keeping us safe and sound, though it has generally been underappreciated for all it does throughout our lives.

Hollywood Is Giving Sentient Robots Top Billing in 2015

On March 6, 2015 a new movie about a sentient robot named "Chappie" will open nationwide. I have not seen this movie yet, but in watching the trailer it is clear that what sets "Chappie" apart from other humanoid robots is that he can learn through sensorimotor experiences and seems to have both the “book smarts” of the cerebrum and the “street smarts” of a cerebellum.

After watching the trailer for Chappie, what intrigues me most from a neuroscientific perspective is that, like a human-child, Chappie the robot needs to go through the classic Piaget stage of sensorimotor development, which has recently been linked strongly to the cerebellum.

This current blog post ballooned in the days since I first encountered Scott Frey's research last month. I have been digging deeper and deeper into recent cerebellar research in an effort to connect the dots and hopefully start a mulit-disciplinary dialogue about how the cerebellum may be pivotal for creating sentient robots and androids. At the end of this post, I've included a bibliography of some of the most relevant science articles I found that address this point.

In addition to Scott Frey's research at Missouri, a different study from February 2015 seems to reflect the potential that cerebellar research holds to do incredibly good things for humanity but also its potential to create a future robotic dystopia. Last month, researchers at RIKEN in Japan reported that they were able to grow Purkinje cells and a cerebellar mass in a petri-dish, using human stem cells.

In The Terminator, Arnold Schwarzenegger’s cyborg had living tissue covering a metal endoskeleton that was designed for combat. It seems possible that future AI may rely on creating artificial living tissue of a human brain that could be grown from stem cells in a laboratory and attached to a synthetic fiber and metal exoskeleton.

In the future, we may have sentient androids with actual brain tissue that has been grown in a laboratory and then attached to a robot or android with superhuman strength and intelligence. (And what would be the pros and cons of inventing such a creature?!)

I wonder how far away we are from having this type of hybrid android becoming a reality? I am not a science-fiction writer, but I can easily imagine Isaac Asimov, Aldous Huxley, Arthur C. Clarke, Daniel H. Wilson, or Philip K. Dick using this latest neuroscience to write a prophetic novel.

The movie Ex Machina is a Sci-Fi thriller, to be released worldwide in April of 2015, about the creation of sentient androids in a screenplay that is likely to put the darker side of AI and the neuroscience of building sentient robots into the spotlight.

Ex Machina explores the complex pyschological landscapes created by AI robots that can think, love, feel, seduce, and hate. Some of the scenes in this trailer remind me of 2001: A Space Odyssey when HAL decides to kill his creators and calmly apologizes for having to annihilate the crew of the spacestation.

For me, researching and advocating for the cerebellum is obviously very personal. One reason this blog post has been difficult for me to write is that for the first time I see a potential backlash for scientists having an advanced understanding of the cerebellum. The cerebellum is so mysterious and elusive. I ponder whether decoding its enigmas might lead to ambiguous breakthroughs that could lead to AI nightmares and future dystopia.

In writing this blog, "roboethics" come to the forefront. What ethical responsibilites do researchers in the neuroscientific community have when it comes to the creation of sentient robots?

During the past couple of years, Google has been acquiring robot companies like Boston Dynamics that are traditionally heavily funded by the military. In 2014, Google bought at least 6 robot companies, and they are being very tight-lipped about their game plan. What is Google's AI agenda?

Traditionally, much of the AI robot research has been subsidized by DARPA's Defense Sciences Office and groups like The Army Research Laboratory’s Army Research Office. On the bright side, all of the federal resources being poured into humanoid robotic development could improve the understanding of how the brain works and lead to treatments that improve human beings' lives throughout their lifespan. But, this research and robotic development could also lead to serious dysfunctions.

The point that Scott Frey and I came back to again and again during our conversation is the importance of constantly striving to adapt and apply these advances in neuroscience and robotics in ways that improve lives.

My main hope is that these studies will end up improving individual lives—like wounded veterans and people with disabilities—while advancing humanity, and reducing the casualties of war.

References: If you'd like to dive deeper into the neuroscience of this topic, below are links to recent journal articles I found while researching this blog post.

1. "Grasping with the press of a button: grasp-selective responses in the human anterior intraparietal sulcus depend on nonarbitrary causal relationships between hand movements and end-effector actions" Frey SH, Hansen M, Marchal N.
2. "Distributed cerebellar plasticity implements generalized multiple-scale memory components in real-robot sensorimotor tasks" Claudia Casellato, Alberto Antonietti et al
3. “Self-Organization of Polarized Cerebellar Tissue in 3D Culture of Human Pluripotent Stem Cells” Keiko Muguruma, Ayaka Nishiyama, Kouichi Hashimoto, Yoshiki Sasai
4. “Adaptive Cerebellar Spiking Model embedded in the control loop: Context switching and robustness against noise” Luque, Garrido et al.
5. “Relationships between regional cerebellar volume and sensorimotor and cognitive function in young and older adults” Bernard JA, Seidler RD.
6. “Cerebellum involvement in cortical sensorimotor circuits for the control of voluntary movements” Rémi D. Proville, Maria Spolidoro et al
7. “Fast convergence of learning requires plasticity between inferior olive and deep cerebellar nuclei in a manipulation task: a closed-loop robotic simulation” Niceto R. Luque, Jesús A. Garrido et al.
8. “Coding of stimulus strength via analog calcium signals in Purkinje cell dendrites of awake mice” Farzaneh Najafi, Andrea Giovannucci, Samuel S-H Wang, Javier F Medina.
9. "Sensory-Driven Enhancement of Calcium Signals in Individual Purkinje Cell Dendrites of Awake Mice" Farzaneh Naja, Andrea Giovannucci,Samuel S.-H. Wang, Javier F. Medina
10. “Realtime cerebellum: A large-scale spiking network model of the cerebellum that runs in realtime using a graphics processing unit” Tadashi Yamazakia, Jun Igarashib
11. “Direct causality between single-Purkinje cell activities and motor learning revealed by a cerebellum-machine interface utilizing VOR adaptation paradigm” Hirata Y, Katagiri K, Tanaka Y.
12. “Robot Learning Manipulation Action Plans by “Watching” Unconstrained Videos from the World Wide Web” Yezhou Yang, Yi Li, Cornelia Fermuller, Yiannis Aloimonos

If you'd like to read more about the cerebellum in layperson terms, please check out my previous Psychology Today blog posts:

• "How Are Purkinje Cells in the Cerebellum Linked to Autism?"
• "Autism, Purkinje Cells and the Cerebellum Are Intertwined"
• "Research Links Autism Severity With Motor Skill Deficiencies"
• “How Is the Cerebellum Linked to Bipolar Disorder?”
• "Is Cerebellum Size Linked to Human Intelligence?"
• "The Neuroscience of Knowing Without Knowing"
• "Primitive Brain Area Linked to Human Intelligence"
• "The Mysterious Neuroscience of Learning Automatic Skills"
• "Too Much Crystallized Thinking Lowers Fluid Intelligence"
• "The Neurobiology of Grace Under Pressure"
• "How Does the Vagus Nerve Convey Gut Instincts to the Brain?"
• "Why Does Overthinking Cause Athletes to Choke?"
• "Toward a New Split-Brain Model: Up Brain-Down Brain"
• "Neuroscientists Discover How Practice Makes Perfect"
• "How Is the Cerebellum Linked to Autism Spectrum Disorders?"
• “Childhood Family Problems Can Stunt Brain Development”
• "The Neuroscience of Calming a Baby"
• “Why Is Dancing So Good For Your Brain?”
• “The Neuroscience of Madonna’s Enduring Success”
• "Gesturing Engages All Four Brain Hemispheres"
• "The Neuroscience of Superfluidity"
• "One More Reason to Unplug Your Television"
• "Better Motor Skills Linked to Higher Academic Scores"
• "Hand-Eye Coordination Improves Cognitive and Social Skills"
• "The Neuroscience of Imagination"
• "Can Practice Alone Create Mastery?"
• "No. 1 Reason Practice Makes Perfect"

Follow me on Twitter @ckbergland for updates on The Athlete’s Way blog posts.

© Christopher Bergland 2015. All rights reserved.

The Athlete’s Way ® is a registered trademark of Christopher Bergland.