The purpose of the research on convergent causal mapping of memory is we don’t really know why some brain lesions or some brain stimulation sites actually cause patients’ memory to get either worse or get better after a lesion or stimulation, respectively. This kind of made me wonder, do we really understand the true localization of memory across the brain? If I was given a patient right now, could I identify where to put an electrode or where to put a stimulation site to make their memory better? I couldn’t...
The purpose of the research on convergent causal mapping of memory is we don’t really know why some brain lesions or some brain stimulation sites actually cause patients’ memory to get either worse or get better after a lesion or stimulation, respectively. This kind of made me wonder, do we really understand the true localization of memory across the brain? If I was given a patient right now, could I identify where to put an electrode or where to put a stimulation site to make their memory better? I couldn’t. So I thought we should go back to the data and investigate across a whole host of diseases and therapeutic stimulations, what was the anatomy associated with memory outcomes? So across four data sets of patients with brain atrophy, strokes, multiple sclerosis, or penetrating traumatic brain injury, as well as four more data sets of deep brain stimulation for epilepsy, Parkinson’s, or kind of stereotactic EEG all over the brain for epilepsy workup, as well as four more data sets of transcranial magnetic stimulation to try and actually make memory better in the context of healthy young adults. We took each of those sites, and we warped them into a common brain space. And despite the fact that these all influenced memory to a variable degree, they very rarely overlapped the same anatomy. They couldn’t cross predict each other and it was very difficult to kind of make sense of why someone got better or worse just based on the location of stimulation or lesion alone. So what we did was we mapped out the brain connections, the brain networks that each of those lesions or stimulation sites were hitting and then we wondered are they actually interacting with kind of a common underlying brain network? And across all 12 of those data sets, independently, we found that they converged upon one similar brain topography. This commonly involved the precuneus, the mesial frontal lobe, the mesial temporal lobe, and the ninth lobule of the cerebellum, as well as a handful of other regions, just the lateral temporal cortex. We found that connectivity of unseen brain lesions or brain stimulation sites to this network was actually able to better explain variance or predict outcomes than other neuroanatomical localizations of memory. So it did a better job than, say, hitting the hippocampus, hitting the circuit of Papez, hitting the default mode network, or, say, lesion network maps of amnesia itself. We then took this and wondered, could it explain some of the heterogeneity in prior studies that had done clinical trials of stimulation for Alzheimer’s disease? If the goal is to localize memory so we have a stimulation target for memory, then it should be at least related to Alzheimer’s. So across 19 studies of TMS and two studies of DBS, we found that just incidentally, the more that their stimulation site hit that convergent memory network, the better the overall outcome of the study was, such that studies that had very, very low connectivity to the convergent memory network actually reported low or sometimes even negative effect sizes, which made Alzheimer’s overall worse, whereas the ones that were highly connected made Alzheimer’s overall better. So the overall next steps of this study are that we hypothesize that we’ve at least found some fundamental network that explains a fundamental aspect of memory, not all of memory and not entirely how memory works across the brain, but some little substrate of memory that’s shared or at least involved across both diseases and brain stimulations when they influence memory. So that is hypothetically our target for brain stimulation trials that want to alleviate symptomatic and all-cause memory dysfunction. So we’re working with transcranial magnetic stimulation expert Joe Taylor at Harvard Medical School to do non-invasive stimulation of this network in patients with Alzheimer’s disease to see if we can make their memory specifically better. And we are going to reprogram patients who have deep brain stimulation electrodes for Alzheimer’s disease sitting inside their fornix to try and make sure the stimulation that’s being delivered by those electrodes maximally hits or engages the convergent memory network. And we’ll see if that can improve their memory performance on the spot. And hopefully this will result in some sort of meaningful target for memory disorders.
This transcript is AI-generated. While we strive for accuracy, please verify this copy with the video.
