🚨 New Preprint 🚨. Targeting intracranial electrical stimulation (ES) to network regions defined within individuals causes network-level effects. By Cyr et al. Q: Can we use individualized network maps from precision fMRI to modulate a targeted network via ES?. A: Yes!. 🧵:

@coryshain.

The most fun image & video creation tool in the world is here. Try it for free in the Grok App.

If you will it to wilt, will it wilt well or willn’t it?.

@bttyeo Finally, a special thanks to co-authors @Chris_Cyr_ @aniaholubecki Lindsay Shi, Maya Lakshman, @SalvoJoe Nathan Anderson, @JamesKragel Sarah Lurie, Joel Voss, Vasileios Kokkinos, @joshuarosenowMD @StephanSchuele Elizabeth Johnson, Christina Zelano.

@bttyeo Overall, we think this lays important groundwork for the use of non-invasive, individualized brain network mapping for targeting intracranial stim to specific networks. This could help improve efficacy of ES, and reduce collateral effects. Here is our proposed framework:

The two types of maps overlapped a lot, but sites where HFES led to network-related behavioral effects were significantly closer to the individualized maps than the group-defined maps:. Sorry, @bttyeo :)

As a bonus analysis, we compared results using individualized vs. group-defined Yeo 17 maps. Note that in some cases it’s hard to pick a matching Yeo network, e.g., “Default B” was the closest network to LANG as defined in individuals:

Note that we had few instances of 1 mA stimulation, so the SPES plot on the left here needs validating in larger samples. But it’s reassuring that the same effect is seen in the HFES plot on the right. More dominance at stim site = more specificity of effects.

When we controlled for those factors, we saw that the Dominance of a given network near the stim site may be an important factor. Targeting stim towards large regions of a given network improves your chances of stimulating only that network:

In sum, our results suggest that there might be a “sweet spot” for network-specific modulation with intracranial ES. For that, stim should be applied:.- at low current intensity (< 1mA).- in white matter, a few mm from the cortex. - within ~10 mm of a network region.

Most of the time, No Effects happened. This may be because few behaviors are actually tested during HFES (e.g., reading, counting, arm movements). The maps could improve future clinical mapping, by showing which networks are at the stim site, and indicating behaviors to test ⚡.

Across all stim sites, HFES led to a behavioral effect related to a given Network N (e.g., language effects for LANG stimulation, Sensorimotor effects for SMOT, etc) more when applied close (e.g., < 5 mm) to that network.

Here are all the instances of clear sensorimotor effects:. Again, most of the time, HFES within 10 mm of the SMOT network led to sensorimotor effects.

Sometimes the stim site that disrupted speech or reading was in a “non-canonical” language region, such as in the ventral temporal cortex or rostral prefrontal cortex, but was still near the language network (LANG, see black outlines):. Note P8 was a right hemisphere implant:

What about behavioral effects?. We saw a similar thing with HFES, which is used for functional mapping (i.e., to evoke/disrupt behaviors such as arm movements, speech arrest) in the clinic. Here are 4 cases, showing all stim sites that led to clear language effects:

Across all sites, we found that stimulation near to a given network, Network N, is more likely to activate that network, even if non-selectively. Stimulating within 10 mm of a network seems to be best, and it then tapers off. (note we grayed out columns with <5 observations)

Here are 3 example stim sites at different distances to the somatomotor (SMOT) network. The sites close to SMOT activated SMOT even at 1 mA, whereas the site far from SMOT did not even at 5 mA.

Taking current and WM depth into account, we began to see subtle effects of the network microenvironment at the stim site. We characterized each stim site based on (1) its distance to each network, and (2) the “Dominance” of each surrounding network:

Note how almost all of the sites where SPES failed to activate any distant network regions (0 on y axis) were within the gray matter (positive values in x axis). Not a good place to stimulate if you want effects!

Building on prior work, we saw that SPES within the white matter is likely to activate more networks. Below are 3 sites that were all stimulated at 5mA:. Importantly, this doesn’t seem to be related to depth into brain: stim in peripheral white matter also has this effect.

Increasing the applied current intensity for SPES resulted in more networks being activated. This suggests that network-specific modulation is more likely to occur at low current intensity (e.g., < ~1 mA). Here is the same site stimulated at increasing mA.