Dr. Strange’s augmented reality surgical assistant

We’re actually done with all of the artifacts from Doctor Strange. But there’s one last kind-of interface that’s worth talking about, and that’s when Strange assists with surgery on his own body.

After being shot with a soul-arrow by the zealot, Strange is in bad shape. He needs medical attention. He recovers his sling ring and creates a portal to the emergency room where he once worked. Stumbling with the pain, he manages to find Dr. Palmer and tell her he has a cardiac tamponade. They head to the operating theater and get Strange on the table.

DoctorStrange_AR_ER_assistant-02.png

When Strange passes out, his “spirit” is ejected from his body as an astral projection. Once he realizes what’s happened, he gathers his wits and turns to observe the procedure.

DoctorStrange_AR_ER_assistant-05.png

When Dr. Palmer approaches his body with a pericardiocentesis needle, Strange manifests so she can sense him and recommends that she aim “just a little higher.” At first she is understandably scared, but once he explains what’s happening, she gets back to business, and he acts as a virtual coach.

DoctorStrange_AR_ER_assistant-08.png

In this role he points at the place she should insert the needle, and illuminates the chest cavity from within so she can kind of see the organ she’s targeting and the surrounding tissue. She asks him, “What were you stabbed with?” and he must confess, “I don’t know.”

DoctorStrange_AR_ER_assistant-11.png

Things go off the rails when the zealot who stabbed him shows up also as an astral projection and begins to fight Strange, but that’s where we can leave off the narrative and focus on everything up to this point as an interface.

Imagine with me, if you will, that this is not magic, but a kind of augmented reality available to the doctor. Strange is an unusual character in that he is both one of the world’s great surgeons and the patient in the scene, so let’s tease apart each.

An augmented reality coach

Realize that Dr. Palmer is getting assistance from one of the world’s greatest surgeons, rendered as a volumetric projection (“hologram” in vernacular). She can talk to him as if he was there to get his advice, and, I presume, even dismiss him if she believes he was wrong. Wouldn’t doctors working in new domains relish the opportunity to get advice from experts until they they have built their own mastery?

Two notes to extend this idea.

In the spirit of evidence-based medicine and big idea, we must admit that it would be better to have diagnoses and advice based on the entirety of the medical record and current, ethical best practices, not just one individual expert. But if an individual doctor prefers to have that information delivered through an avatar of a favored mentor, why not?

The second note to anyone thinking of this as a real world model for an AR assistant: I would expect a fully realized solution to include augmentations other than just a human, of course, such as ideal angles for incisions, depth meters, and life signs.

A (crude) body visualization

One of the challenges surgeons have when working with internal damage is that the body is largely opaque. They have to use visualization tools like radiographs and (very) educated guesses to diagnose and treat what’s going on inside these fleshy boxes of ours. How awesome that the AR coach can help illuminate (in both senses) the body to help Dr. Palmer perform the procedure correctly?

DoctorStrange_AR_ER_assistant-13.png

Admittedly, what we see in Doctor Strange is a crude version. This same x-ray vision appeared with more clarity and higher resolution in two other films, as cited in the Medical chapter of Make It So. In Lost in Space, the medical table projects a real-time volumetric scan of the organs inside Judy’s body into the eyes of the observers.

7809592870_cf8f3fdccf_o.jpg

In Chrysalis, Dr. Bruger sees a volumetric display of the patient on which she is teleoperating.

7809609070_b65b6fb936_o.jpg

But despite its low-resolution, I wanted to draw it out as another awesome and somewhat subtle part of the way this AR assistant helps the doctor.

A queryable patient avatar

Lastly, consider that Dr. Palmer is able to ask her patient what happened to him. Of course in the real world passed out patients aren’t able to answer questions, but of course understanding the events that led to a crisis are important. I can imagine several sci-fi ways that this information might be retrievable from the world.

  • Trace evidence on the patient’s body: High-resolution sensors throughout the operating theater could have automatically run forensic analysis on the patient the moment they entered the room to determine type of wound and likely causes, such as  microscopic detection of soot in entrance wounds.
  • Environmental sensors: If the accident happened in a place with sensors that are queryable, then the assistant could look at video footage, or listen in to microphones in the environment to help piece together what happened. Of course the notion of a queryable technological panoptican has massive privacy issues which cannot be overlooked, but if the information is available to medical professionals, it would be tragic to ignore it in genuine crises.
  • Human witnesses can provide informative narratives. Witness and first responders may be on record already. But in looking at the environmental sensors, the assistant might be able to instantly reach out to those who have not. Imagine one of these witness, shaken by the event he saw, on a commute home. His phone buzzes and it is the assistant saying, “Hello, Mr. Mackinnon. Records indicate that you were witness to a violent crime today, and your account of the event is needed for the victim, who is currently in surgery. Can you take a moment to answer some questions?”
  • Patient preferences should be automatically exposed and incorporated via the assistant as well. If the patient was a Jehovah’s Witnesses, for instance, then their desire not to have a blood transfusion should be raised in whatever form the assistant takes.

An surgical assistant could automatically query all of these sources to make a hypothesis of what happened and advise the procedure. This could be available doctor for the asking, volunteered by the assistant at a lull in more critical action, or offered by the assistant as a preventative. I suspect it’s more likely the doctor would ask the assistant than the patient, e.g. “OK, ERbot, what happened to this guy?” but if the doctor prefers, she should be able ask in the second person, as Dr. Palmer does in the scene, and the system should reply appropriately.

Sure, in this context, it’s magic, but since we can imagine how it could be done with technology, this scene gives us a very dense set of inspirational ideas for the future of surgical assistants.

Zorg bomb

TheFifthElement-ZorgBomb-001

When Zorg believes he has recovered the sacred stones, he affixes a bomb to the door of Plavalaguna’s suite. The bomb is a little larger than a credit card, with a slot at the top for a key card to be dropped in. The front of the bomb houses all the buttons and lights. The bottom and top edges are rounded back.

The interface for the bomb is quite simple. Zorg presses three large, transparent buttons along the top in order from left to right to activate the bomb. These buttons glow bright red during the countdown. Below these buttons, four red LEDs blink in succession counting off quarter seconds. At the bottom of the display a 4-character, 7-segment timer counts down from the time set: 20 minutes. The device audibly ticks off each second as it passes.

timebomb

Activation

An (adhesive? magnetic?) backing lets Zorg simply place the bomb on the wall to affix it there. Zorg presses the three large buttons in order from left to right to activate it and start the countdown.

Activation analysis

The bomber is after simple activation, but also wants very much to avoid accidental activation. Pressing the buttons in order might happen accidentally, for example from a tire or foot rolling across it. Better would be to have the activation code something much less likely to happen accidentally, like 1-3-2 or 2-3-1.

There’s also a question of whether a bomber would put giant glowing lights, reflective yellow tape, or an audible tick on the bomb (LEDs, if you didn’t know, don’t come with a ticking sound built in.) Each of these draws attention to the bomb, giving helpless victims time to evacuate, alert the authorities, or inform any explosive ordnance disposal personnel that happen to be wandering by. Yes, Zorg wants the bomb to explode, but only after a certain time, so he can get away. He should affix the bomb in some hidden place and design it with a less attention-getting display to suit his fiendish goals.

TheFifthElement-ZorgBomb-003

Deactivation

Once Zorg realizes that the box he stole was empty, he returns to the Fhloston Paradise liner to look for the stones. His first task is to deactivate the bomb. To do this he pulls out a keycard, and gingerly holds it above the bomb. His caution and nervousness implies that it has a jostle-sensitive anti-handling sensors, and that if he bumped it, it would go off. Fortunately for him, he manages to slip the card in without jostling the bomb, and sure enough, it stops with five seconds to go.

TheFifthElement-ZorgBomb-006

Deactivation analysis

The keycard is a mostly-smart deactivation strategy. As we can see, Zorg is quite nervous during the deactivation, and in such high-stress times, it’s better to rely on an object than a stressed villain’s memory for something like a password. The card is thin like a credit card and can fit in a wallet, so it’s easy to carry around. There’s a risk that the card could be misplaced, but the importance of the key will ensure that Zorg will keep track of it. There’s a risk it could be ruined and become useless, but we can presume Zorg made it with tough, ruggedized materials.

The problem with the shape is one of orientation. There are four ways a card can be oriented to a slot, and looking at the card, there is no clear indication of the correct one. The copper circuitry printed on both sides is asymmetrical, so it’s at least possible to tell the current orientation. Perhaps this is the “password” that the system requires, and the random stranger picking it up only has a one in four chance of getting it right.

Fortunately for Zorg, he remembers the correct orientation, and is able to stop the bomb.

TheFifthElement-ZorgBomb-007

Or, this bomb, anyway.

MedPod

Early in the film, when Shaw sees the MedPod for the first time, she comments to Vickers that, “They only made a dozen of these.” As she caresses its interface in awe, a panel extends as the pod instructs her to “Please verbally state the nature of your injury.”

Prometheus-087

The MedPod is a device for automated, generalized surgical procedures, operable by the patient him- (or her-, kinda, see below) self.

When in the film Shaw realizes that she’s carrying an alien organism in her womb, she breaks free from crewmembers who want to contain her, and makes a staggering beeline for the MedPod.

Once there, she reaches for the extended touchscreen and presses the red EMERGENCY button. Audio output from the pod confirms her selection, “Emergency procedure initiated. Please verbally state the nature of your injury.” Shaw shouts, “I need cesarean!” The machine informs her verbally that, “Error. This MedPod is calibrated for male patients only. It does not offer the procedure you have requested. Please seek medical assistance else–”

Prometheus-237

I’ll pause the action here to address this. What sensors and actuators are this gender-specific? Why can’t it offer gender-neutral alternatives? Sure, some procedures might need anatomical knowledge of particularly gendered organs (say…emergency circumcision?), but given…

  • the massive amounts of biological similarity between the sexes
  • the needs for any medical device to deal with a high degree of biological variability in its subjects anyway
  • most procedures are gender neutral

…this is a ridiculous interface plot device. If Dr. Shaw can issue a few simple system commands that work around this limitation (as she does in this very scene), then the machine could have just done without the stupid error message. (Yes, we get that it’s a mystery why Vickers would have her MedPod calibrated to a man, but really, that’s a throwaway clue.) Gender-specific procedures can’t take up so much room in memory that it was simpler to cut the potential lives it could save in half. You know, rather than outfit it with another hard drive.

Aside from the pointless “tension-building” wrong-gender plot point, there are still interface issues with this step. Why does she need to press the emergency button in the first place? The pod has a voice interface. Why can’t she just shout “Emergency!” or even better, “Help me!” Isn’t that more suited to an emergency situation? Why is a menu of procedures the default main screen? Shouldn’t it be a prompt to speak, and have the menu there for mute people or if silence is called for? And shouldn’t it provide a type-ahead control rather than a multi-facet selection list? OK, back to the action.

Desperate, Shaw presses a button that grants her manual control. She states “Surgery abdominal, penetrating injuries. Foreign body. Initiate.” The screen confirms these selections amongst options on screen. (They read “DIAGNOS, THERAP, SURGICAL, MED REC, SYS/MECH, and EMERGENCY”)

The pod then swings open saying, “Surgical procedure begins,” and tilting itself for easy access. Shaw injects herself with anesthetic and steps into the pod, which seals around her and returns to a horizontal position.

Why does Shaw need to speak in this stilted speech? In a panicked or medical emergency situation, proper computer syntax should be the last thing on a user’s mind. Let the patient shout the information however they need to, like “I’ve got an alien in my abdomen! I need it to be surgically removed now!” We know from the Sonic chapter that the use of natural language triggers an anthropomorphic sense in the user, which imposes some other design constraints to convey the system’s limitations, but in this case, the emergency trumps the needs of affordance subtleties.

Once inside the pod, a transparent display on the inside states that, “EMERGENCY PROC INITIATED.” Shaw makes some touch selections, which runs a diagnostic scan along the length of her body. The terrifying results display for her to see, with the alien body differentiated in magenta to contrast her own tissue, displayed in cyan.

Prometheus-254

Prometheus-260

Shaw shouts, “Get it out!!” It says, “Initiating anesthetics” before spraying her abdomen with a bile-yellow local anesthetic. It then says, “Commence surgical procedure.” (A note for the grammar nerds here: Wouldn’t you expect a machine to maintain a single part of speech for consistency? The first, “Initiating…” is a gerund, while the second, “Commence,” is an imperative.) Then, using lasers, the MedPod cuts through tissue until it reaches the foreign body. Given that the lasers can cut organic matter, and that the xenomorph has acid for blood, you have to hand it to the precision of this device. One slip could have burned a hole right through her spine. Fortunately it has a feather-light touch. Reaching in with a speculum-like device, it removes the squid-like alien in its amniotic sac.

OK. Here I have to return to the whole “ManPod” thing. Wouldn’t a scan have shown that this was, in fact, a woman? Why wouldn’t it stop the procedure if it really couldn’t handle working on the fairer sex? Should it have paused to have her sign away insurance rights? Could it really mistake her womb for a stomach? Wouldn’t it, believing her to be a man, presume the whole womb to be a foreign body and try to perform a hysterectomy rather than a delicate caesarian? ManPod, indeed.

Prometheus-265

After removing the alien, it waits around 10 seconds, showing it to her and letting her yank its umbilical cord, before she presses a few controls. The MedPod seals her up again with staples and opens the cover to let her sit up.

She gets off the table, rushes to the side of the MedPod, and places all five fingertips of her right hand on it, quickly twisting her hand clockwise. The interface changes to a red warning screen labeled “DECONTAMINATE.” She taps this to confirm and shouts, “Come on!” (Her vocal instruction does not feel like a formal part of the procedure and the machine does not respond differently.) To decontaminate, the pod seals up and a white mist fills the space.

OK. Since this is a MedPod, and it has something called a decontamination procedure, shouldn’t it actually test to see whether the decontamination worked? The user here has enacted emergency decontamination procedures, so it’s safe to say that this is a plague-level contagion. That’s doesn’t say to me: Spray it with a can of Raid and hope for the best. It says, “Kill it with fire.” We just saw, 10 seconds ago, that the MedPod can do a detailed, alien-detecting scan of its contents, so why on LV-223 would it not check to see if the kill-it-now-for-God’s-sake procedure had actually worked, and warn everyone within earshot that it hadn’t? Because someone needs to take additional measures to protect the ship, and take them, stat. But no, MedPod tucks the contamination under a white misty blanket, smiles, waves, and says, “OK, that’s taken care of! Thank you! Good day! Move along!”

For all of the goofiness that is this device, I’ll commend it for two things. The first is for pushing the notion forward of automated medicine. Yes, in this day and age, it’s kind of terrifying to imagine devices handling something as vital as life-saving surgery, but people in the future will likely find it terrifying that today we’d rather trust an error prone, bull-in-a-china-shop human to the task. And, after all, the characters have entrusted their lives to an android while they were in hypersleep for two years, so clearly that’s a thing they do.

Second, the gestural control to access the decontamination is well considered. It is a large gesture, requiring no great finesse on the part of the operator to find and press a sequence of keys, and one that is easy to execute quickly and in a panic. I’m absolutely not sure what percentage of procedures need the back-up safety of a kill-everything-inside mode, but presuming one is ever needed, this is a fine gesture to initiate that procedure. In fact, it could have been used in other interfaces around the ship, as we’ll see later with the escape pod interface.

I have the sense that in the original script, Shaw had to do what only a few very bad-ass people have been willing to do: perform life-saving surgery on themselves in the direst circumstances. Yes, it’s a bit of a stretch since she’s primarily an anthropologist and astronomer in the story, but give a girl a scalpel, hardcore anesthetics, and an alien embryo, and I’m sure she’ll figure out what to do. But pushing this bad-assery off to an automated device, loaded with constraints, ruins the moment and changes the scene from potentially awesome to just awful.

Given the inexplicable man-only settings, requiring a desperate patient to recall FORTRAN-esque syntax for spoken instructions, and the failure to provide any feedback about the destruction of an extinction-level pathogen, we must admit that the MedPod belongs squarely in the realm of goofy narrative technology and nowhere near the real world as a model of good interaction design.