[Eeglablist] Source localization and hippocanpus

[Scott Makeig]

An important factor not yet addressed in this thread concerning imputing
hippocampal origins to EEG sources is the ambiguity in EEG source *depth*
estimation produced by using a template (one-value-fits-all) value for
skull conductance (e.g., relative to cortical tissue conductance). In 2013,
Zeynep Akalin Acar
<https://urldefense.com/v3/__https://link.springer.com/content/pdf/10.1007/s10548-012-0274-6.pdf__;!!Mih3wA!ARxtv65vBCQj3uGqJxF-izHH1MAPWsRLLZ5kAK043QoxC1kaT2BDBCuG_QnxoOtnOKNLKSkPNhtykTnejW_v$ >
showed that misestimation of this key value could  move the estimated
dipole depth of a simulated source inwards|outwards by up to 2 cm.

Subsequent analyses using our SCALE algorithm
<https://urldefense.com/v3/__https://www.sciencedirect.com/science/article/pii/S1053811915007442?casa_token=TNbNtkbte48AAAAA:H_y-zGCj6GaVzozPW1Qz1S2hNrlt0AZzMkhbBGS-sxHWYbm46ChN59POMXWoFf00KboVJRx-_Q__;!!Mih3wA!ARxtv65vBCQj3uGqJxF-izHH1MAPWsRLLZ5kAK043QoxC1kaT2BDBCuG_QnxoOtnOKNLKSkPNhtykeFilC6y$ >
(2016)
suggested that skull conductance of adults in psychophysiological EEG
experiments varies by up to 3:1 or more (from ~20 brain/skull conductivity
ratio, up to ~75; in young children, ~8; still lower in infants). Fitting a
source projection map with a deep equivalent dipole (often a result for
independent component 'brain source' maps ),using either a template head
model OR an individual model constructed using a template skull-conductance
value -- may therefore be misleading.

That is, because of possible misestimation of skull conductance in
template-based models, the error ellipse for estimated equivalent-dipole
source location is not shaped like an orange, but more like a radially
oriented zuccini (courgette) ...

The SCALE algorithm* enables stable estimation
<https://urldefense.com/v3/__https://ieeexplore.ieee.org/iel7/9870821/9870822/09872004.pdf?casa_token=eawNGHoZPoEAAAAA:rhwfqaClncG8UfR9CqVs20aVb4fda8AGGDmZQSWuGTMYHwsh9C0w6F4wBbIDORYhSH34GwvwFw__;!!Mih3wA!ARxtv65vBCQj3uGqJxF-izHH1MAPWsRLLZ5kAK043QoxC1kaT2BDBCuG_QnxoOtnOKNLKSkPNhtykc6JLBWF$ >*
of individual participant skull conductivity from any (decent) EEG
recording and structural participant MR head image, though unfortunately a
tool for it is not yet available.

Scott Makeig


> On Thu, Jan 15, 2026 at 1:01 PM Rawls, Eric via eeglablist <
eeglablist at sccn.ucsd.edu> wrote:

> I perhaps have a different opinion on this than some of those who have
> responded, but in my own reading of the literature, hippocampus has always
> seemed a more plausible solution for source generator of scalp recorded
> event related potentials than other subcortical structures.
>
> For example, the basal ganglia do not have the required cellular
> architecture, that is a laminar organization with arrayed pyramidal neurons
> and dendrites which are aligned spatially to produce synchronous
> postsynaptic potentials.
>
> Prior results implicating the basal ganglia in the generation of, for
> example, the reward positivity are more easily explained by a superposition
> of two sources in the brain, which my current research suggests being
> medial frontal cortex and posterior cingulate cortex. The same argument was
> made in a rebuttal to that paper, in which a simulation showed that the
> same pattern could be easily explained by a two-generator motif.
>
> On the other hand, the hippocampus has the necessary physiology of
> spatially aligned pyramidal cells which can form synchronous potentials
> projecting to the scalp surface. This is according to simulation and
> empirical recordings, for example, depth recordings from hippocampus
> simultaneous with ecog grid recordings. In this regard, hippocampus is
> physiologically somewhat unique, and in fact, might be better classified as
> cortical tissue rather than subcortical based on some of these features.
>
> Best,
> Eric Rawls
>
>
>

-- 
Scott Makeig, Research Scientist and Director, Swartz Center for
Computational Neuroscience, Institute for Neural Computation, University of
California San Diego, La Jolla CA 92093-0559, http://sccn.ucsd.edu/~scott