Science Behind the Character — Karma

Science Behind the Character: Karma Karma can take control of somebody else's body, use their senses, and make them act on her decisions. That sounds straightforward until you stop treating the body like a vehicle with a spare seat. The nervous system controlling that body grew up with it. Its movements, memories, expectations and senses have been developing together for years. Taking it over would require far more than sending a stronger signal. I'm continuing through the New Mutants with Karma, now named Xuân Cao Mạnh in the comics, with Xi'an Coy Manh appearing throughout the older material. Her defining ability is psychic possession. The classic description includes sensory access, difficulty adjusting to an unfamiliar body, and divided attention when she controls several people. Those limitations are useful because they give the power something we can actually analyze. The [older Karma reference](https://www.marveldirectory.com/individuals/k/karma.htm) describes those features; the [updated character index](https://uncannyxmen.net/glossary/characters/karma) records the later name correction and broader psychic abilities. The question I'm interested in is how much information she would need to exchange with another nervous system, and which parts of that nervous system she would have to leave working. Start by separating three jobs: detecting another person's neural activity, understanding it, and changing it in a useful way. Real neuroscience can accomplish limited versions of these jobs with instruments. Combining them into effortless, remote control of a stranger is firmly fictional. A brain doesn't transmit its decisions in a universal language. Electrical activity in one person's motor cortex isn't a command packet labeled LEFT HAND, CLOSE FINGERS. Neurons participate in populations whose activity depends on the movement, context, training, sensory feedback and individual anatomy. Even an accurate recording needs interpretation. That matters for possession because Xuân would need to identify where relevant information is represented in a brain she has never encountered. A person with different proportions, an old injury or an unfamiliar movement skill could require a different control strategy. Knowing the intended action would be only the beginning. There is a fascinating real comparison in a 2014 experiment by Rajesh Rao and colleagues. One participant's EEG activity was interpreted by a computer and used to trigger magnetic stimulation over another participant's motor cortex, producing a hand movement. The apparatus linked a narrow task-related signal to a simple response. Both people were participating in an arranged experiment with equipment. Nobody acquired another person's memories or took over their life. [The original brain-to-brain interface paper](https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0111332) is worth reading precisely because the actual achievement is more specific than the headline sounds. Karma would need that general idea expanded in almost every direction: vastly richer sensing, precise targeting, interpretation across different brains, continuous sensory return, and access without attached equipment. We have no demonstrated biological mechanism for that package. Simply saying the brain is electrical doesn't solve the distance problem. Electrodes normally need to be on or near the body, and recording outside the skull blends the activity of large populations. An external magnetic field can affect neural activity under controlled conditions, but broad excitation doesn't contain instructions for a coordinated conversation or a careful walk down stairs. The precision is the hard part. Reading meaningful content adds another layer. In 2023, Jerry Tang and colleagues reconstructed aspects of language meaning from fMRI recordings using models trained with individual participants. Their decoder could produce descriptions resembling the meaning of a story without reproducing every exact word. It needed substantial subject-specific data and cooperation in that experiment. This is a useful demonstration of a relationship between brain activity and meaning, not a technique for silently reading an unwilling stranger across a room. [Their study explains both the achievement and its privacy limits.](https://www.nature.com/articles/s41593-023-01304-9) For a fictional Karma model, I'd therefore give her an unknown interaction that can locate and communicate with neural systems. I'd keep that explicitly fictional. Then I'd use actual neuroscience to ask what the interaction must accomplish once it connects. Movement gives us the clearest test. Suppose Xuân wants a possessed person to pick up a glass. She needs the hand to arrive at the right place with the right orientation. The fingers have to close with enough force to prevent slipping without breaking the glass. The shoulder, elbow and wrist have to cooperate, and the body must stay balanced as its weight shifts. Ordinarily, the person doesn't consciously calculate every muscle contraction. The nervous system uses learned control, prediction and incoming sensory information. Touch reports contact. Proprioception reports aspects of the position and movement of the body. Vision helps estimate where the object is. Those streams continuously revise the action. The most economical fictional explanation would let Karma issue higher-level goals while borrowing the host's existing movement machinery. Otherwise she has to manage an enormous number of details that the host already knows how to handle. Asking the body to reach for the glass is a more manageable task than individually commanding every muscle involved. That model also explains why an unfamiliar body could initially feel awkward. Her expected result and the actual result wouldn't quite agree. Longer arms, different strength, altered balance or a different center of mass change what the same intention accomplishes. She would need to learn the relationship between command and outcome while the action is happening. A real clinical comparison comes from a 2017 study in which an implanted brain-computer interface helped a person with tetraplegia control electrical stimulation of his own arm and hand muscles. The system restored useful reaching and grasping in the study setting, with equipment and arm support. It demonstrates how neural intent can be connected to movement through an engineered pathway. It also shows that muscle recruitment, calibration and the mechanics of the limb remain essential. [Ajiboye and colleagues' proof-of-concept study](https://pubmed.ncbi.nlm.nih.gov/28363483/) provides the details. Karma's version would have to calibrate itself quickly enough to function in a fight. The comic-book adjustment period is a sensible place for that difficulty to appear. Feedback delay would matter too. Imagine a hypothetical connection that adds a tenth of a second between a change in the host's position and Karma receiving that information. At a walking speed of 1.5 meters per second, the body travels 15 centimeters during that delay. At a faster 10 meters per second, it travels a full meter. Those are illustrative calculations, not established limits on her power. She could compensate partly by predicting where the body is going, much as normal movement uses prediction. Unexpected events would still create errors. A slippery step or a sudden impact changes the situation before the delayed information arrives. That makes local reflexes valuable. Her control would work better if the host could automatically correct balance and posture without waiting for every instruction to complete a round trip. Possession therefore shouldn't require every ordinary brain function to shut down. Breathing, circulation, posture and sensory processing have to remain coordinated. If Xuân suppresses the host's ability to choose actions, that doesn't imply she has switched off all the machinery keeping the person alive. The older reference compares the host's reduced awareness to dreamless sleep, but a narrative comparison isn't an electroencephalogram or a clinical diagnosis. There is a separate question about strength. A possessed body still has the mechanical limits of its muscles, tendons, bones and joints. Changing who gives the orders doesn't automatically add stronger tissue. Exertion also needs oxygen, fuel and heat removal. A control power might prevent someone from responding normally to discomfort, but ignoring warning signals doesn't make the physical damage disappear. This also tells us why the energy budget is subtler than it looks. Karma doesn't necessarily have to supply all the energy used by another person's muscles. The host can burn its own metabolic fuel. Her contribution could be control information plus whatever energy the fictional connection requires. A small command can direct a much larger existing source of energy. That is ordinary engineering logic, although the remote psychic interface remains imaginary. Controlling several bodies raises an attention problem before it becomes an energy problem. For a deliberately simple bookkeeping example, suppose a controller tracks 20 movement variables per person, updates each 50 times per second, and represents each update with 12 bits. That's 12,000 bits per second per body before sensory feedback, timing, error correction or any representation of thoughts. Ten bodies would make it 120,000 bits per second in that toy system. Those numbers are not an estimate of the brain's actual bandwidth, and they aren't a lower bound on possession. The encoding choices are arbitrary. The point is that detailed supervision grows as you add independently moving bodies. A broad instruction to ten people can be compact. Keeping ten people in precise coordination while each encounters different obstacles is a much larger task. That makes the classic divided-attention limitation meaningful. If Xuân focuses on one host, another might keep following a simple instruction but respond poorly to something unexpected. She could either switch attention rapidly or maintain several partly independent control processes. The latter would imply an extraordinary cognitive architecture, in addition to the connection itself. Receiving somebody else's senses creates another difficult problem: keeping track of whose information belongs to whom. If Xuân sees one room through her own eyes and a different room through a host's eyes, the perspectives cannot simply be added together. Each has its own location, body orientation and possibilities for action. Otherwise a threat approaching the host might provoke a movement in the wrong body. I'd expect a coherent fictional system to label those streams, filter most of them, and bring selected information into attention. Full conscious access to every sight, sound, pressure and movement from a crowd would be a different ability from giving that crowd commands. Memory and identity are where I would be especially careful about the claims. Moving a hand doesn't establish access to an autobiography. Disrupting awareness doesn't establish the ability to erase a particular memory. Retrieving information doesn't establish that it can be copied accurately into another brain. These are different operations on overlapping but distinct systems. Memories also aren't necessarily accurate records of events. A psychic character could access what somebody remembers without learning what objectively happened. Beliefs, mistakes and imagined scenarios still have neural representations. If Karma retrieves a vivid recollection, its vividness alone wouldn't make it true. The comics can treat a personality or psychic presence as something that survives transfer between minds. Real neuroscience hasn't demonstrated detachable persons that can be moved into another brain. We can discuss changes in experience, action and memory without pretending we've solved that much larger question. The ethical implications follow directly from the mechanics. A body moving under outside control hasn't consented just because it appears to cooperate. Restoring the person's ability to act matters independently of whether the controller intended a good outcome. That makes restraint an important part of how Karma uses the ability, rather than a decorative concern added after the science. My most consistent model of her power would be a two-way neural interface with fast adaptation, selective suppression of voluntary control, access to the host's existing movement systems, and a strict limit on how much she can consciously supervise. It would preserve ordinary physiology, keep sensory streams distinct and separate motor control from memory access. The impossible part is the interface and its precision. The interesting science begins with everything it has to coordinate. Karma's power is a reminder that a human body comes with a lifetime of learned control. Taking command of it without destroying that organization would be an astonishing feat, even before she tries to control a second one.