Due to the blood-brain barrier, it is quite difficult to get substances injected into the blood into the brain (though some obviously do, otherwise there would be no psychoactive substances). Injecting into brain tissue directly is possible but probably not useful in an emergency: if you want to inject into cortex directly, you have to inject at ulta-low (yet positive) pressures (otherwise you destroy a lot of tissue just by injecting), and diffusion from the site of injection would be slow. In other words, you would have to drill a thousand holes into someones skull, and inject tiny volumes over the course of minutes each time. You could inject into the ventricles. That can be done quicker and with larger volumes without incurring too much damage, but again, lymphatic flow is very slow, so getting the drug into the tissue would take a long time (hours). So from first principles, you are fighting an uphill battle, and the incline is very steep.

The findings in the paper are interesting in the latter scenario, as they identify the deep cortical layers as the likely drivers of the wave of death (not a surprising finding for various reasons so quite likely true). Deep cortical layers are closer to the ventricles, so access would be a bit better than to the cortex in general.