The Cycle Just Stops

On October 4, 2021, UCSF announced the results of a study in which a closed-loop neurostimulation device had been implanted in a 36-year-old woman identified only as Sarah, whose depression had resisted every standard treatment for years. The device, described that month in Nature Medicine by Katherine Scangos and her colleagues, watches for one specific signal, a burst of gamma-frequency activity in the amygdala that tracks her most severe depressive states. When the pattern appears, it sends a six-second pulse of stimulation into a different part of her brain. It doesn’t wait for Sarah to notice she is sliding. It fires many times a day, and every pulse is set off by a signal in her brain, not by anything she noticed, decided or did.

Before the implant, Sarah’s description of her life was bleak. “I was at the end of the line,” she said in UCSF’s announcement. “I was severely depressed. I could not see myself continuing if this was all I’d be able to do… It was not a life worth living.” After the device was turned on, the change came fast enough to surprise her. “Within a few weeks the suicidal thoughts just disappeared,” she told CBS News. “My lens on the world just changed.”

The interesting part isn’t that the treatment worked. It’s how Sarah explains getting better, because in back-to-back paragraphs of the same announcement she gives two different answers. First she remembers the speed: “the lessening of the depression was so abrupt, and I wasn’t sure if it would last.” Then, once it had lasted, she gives the credit to herself: “I’ve come to realize that the device really augments the therapy and self-care I’ve learned while being a patient here at UCSF.” In that version the work is hers, years of it, and the implant is a helper. In the next paragraph, describing what happens when a depressive spiral starts, she reaches for something with no work in it at all: “those thoughts still come up, but it’s just… poof… the cycle stops.” Nobody describes therapy they worked hard at as a poof. She doesn’t seem to notice she has described both.

Both descriptions are honest, and both probably report something real. The problem is that they can’t be the same kind of event. “Augments the therapy I’ve learned” describes a tool she uses, the way a notebook helps memory only when someone picks it up and writes in it. “Poof, the cycle stops” describes an intervention set off by a gamma-band signature in her amygdala, not by anything her attention caught or her will chose, which is how the closed loop in the paper is designed to work. In one picture Sarah is using a tool. In the other, a circuit is doing something to Sarah on a trigger her awareness has no part in. Her words move between the two without a visible seam.

The psychiatrist Peter Kramer raised a close cousin of this worry in 1993, in Listening to Prozac, about an older technology. His patients often said the drug let them finally feel like themselves, the person they had always meant to be. Kramer, watching the same change from the prescribing side, coined “cosmetic psychopharmacology” for what looked less like a self restored than a personality adjusted to order. Ten years later, in Better Than Well (2003), the philosopher Carl Elliott argued that psychiatry had lost the ability to tell those two descriptions apart, because both can be honest reports of the same relief, told from two different places in one person’s account of her own mind. An implant that responds to a brain signal rather than a conscious decision makes that old problem sharper, not easier. Prozac took weeks to work, and the patient’s own reflection kept running alongside it the whole time. Sarah’s device is built, on purpose, so that her reflection is never what sets it off.

None of this argues against the treatment. A life that was not worth living became one that is, by Sarah’s own account, and for once the free will and determinism debate lands on something real instead of a thought experiment about levers and neurons that never touched anyone’s depression. But a good outcome doesn’t make the question go away. Sarah’s two descriptions, offered side by side in one announcement, describe an event that was at once something she did and something done to her, and there is no experiment left to run that could tell her, or anyone reading her words later, which of her two sentences was the truer report of what happened in the moment the cycle stopped.