The literature I've seen indicates that you may get exactly the opposite effect i.e. neurogical impairment at this wavelength. I can't remember the source but I found it on the internet and I believe it might have been in association with the original paper on mice and memory enhancement.

The 1070 nm has little to do with penetration through the skin and bone of the skull. Bright infrared can penetrate skin and bone easily as a simple experiment with shooting a television remote control through your hand will demonstrate. (channels change) (See more on usability of TV remotes below)

I suspect that a few things are going on with the wavelength. 1070 is very low in energy and there may be a significant difference between ionization energies for some cells or particular atoms in structures in cells that are being affected by the illumination. 940 may ionize 1070 may not. There may be a thin line between harm and help.

Why might illumination of the brain-- an organ that lives in the dark--stimulate the production of neural connections? Is the illumination simulating other internal stimulation that naturally stimulates neural regeneration in the brain? One might observe that one of the few places on the human body that maintains hair is the head and that hair may block light from the brain. Maybe a way of metering the amount of light that gets to the brain. Neurons eventually evolved into retinal cells so we can assume some inherent photosensitivity in neurons.

I also wonder whether a constant light source might not be less effective than a modulated one. Most natural neural stimulation is modulated. Nerves turn on and off in rhythm and in response to external stimulation. The impression that I've gotten from the articles is that the light source is constant. Modulation also allows one to drive the LEDs with more current hence more brightness without heatsinking. An easy modulator can be built with a TV remote control --available at a dollar store. Solder one switch closed and put your 1070 in place of the one in the control and you are ready to roll. TV remotes also overdrive the LEDs to increase remote control range. There are two major frequency types in the remote signal--a low frequency component for the packet of information and the higher frequency that has the digitally encoded information that tells the TV what the actual command is. The lower frequency is in the range of natural brain rhythms. Just a few thoughts for those experimenters among you.