Columbia says depression stalls new neurons. The subtype it implies has no test yet

Pattern separation is the load-bearing idea here, and in humans it is assembled rather than demonstrated. The hippocampus lets similar memories be held apart and keeps the feeling attached to one from bleeding into another; when that fails, a friend who is quiet over lunch gets filed with every earlier experience of rejection, which is how Maura Dupont, the Columbia psychiatrist who led the work, describes patients who can retrieve only the negative version of their own history ^[6][2]. Mouse work established that adult neurogenesis is necessary for that sorting ^[10]. The nearest human evidence is a cohort of brain tumour patients whose hippocampal neurogenesis was eliminated by radiation, which the group reads as suggesting the relationship carries over ^[11]. Count the links and you get one animal necessity result and two human observations, with no experiment that manipulates neurogenesis in a depressed person ^[9]. Dupont says the mechanism is not fully worked out, particularly in humans, and that newborn neurons appear to help because they take up new experiences readily ^[1].

That hedge is not decoration. Her treatment sentence carries its own condition: switching neurogenesis back on may help "in some people" by rewiring the hippocampal circuit ^[17]. So the subtype argument sits inside the finding rather than being appended to it, and it arrives with an awkward requirement. Nothing in the material describes a way to measure adult hippocampal neurogenesis in someone who is alive ^[13]. A molecular subtype you cannot detect in a patient is a research category, not a prescribing category.

Scale is worth keeping in view. Most of the roughly 100 billion neurons a person has were made before birth, and the adult hippocampus adds a small number on top of that ^[4]. The claim is that this thin stream of cells carries disproportionate weight for resilience, and it lands against a history in which depression was read as a shortage of neurotransmitters, serotonin above all. Dupont's position is that the disorder instead comes from several problems in how neurons adapt to stress and to a changing environment ^[3]. Only a few regions of the adult brain make new neurons at all, which is part of why the hippocampus keeps absorbing the attention ^[5].

What Columbia has produced is a human data point aimed the way the rodent literature already pointed, and it is labelled a first ^[15][10]. What the field still does not have is any way to tell which patient in the waiting room is the one whose hippocampus has stopped ^[13].