Skip to main content
. 2009 Oct 7;102(6):3554–3572. doi: 10.1152/jn.00535.2009

Table 3.

Summary of MEG results, model reproductions, and predicted neural mechanisms

MEG Data Model Reproduction Predicted Neural Mechanisms
(1) Mu rhythm originating from SI with mu-alpha and mu-beta components (Fig. 3). Net subthreshold intracellular current flow in layers II/III and V PNs (the simulated SI MEG signal) produced a mu rhythm with mu-alpha and mu-beta components (Fig. 6C). The mu-alpha rhythm is generated by thalamic lemniscal input generating current flow propagation up the PN dendrites primarily away from the soma. The mu-beta rhythm is generated by input from other cortical regions outside the hand representation expressing a mu-alpha oscillation. This input arrives in layers II/III and generates current flow “down” PN dendrites toward the soma. The arrival time of this intracortical input is stochastic and nearly synchronous in alternating with the ongoing approximately 10-Hz FF thalamic input. Alternative distal dendritic inputs (e.g., from nonspecific thalamic projections) could also contribute to mu-beta emergence.
(2) Mu rhythm oscillates around zero with both current polarities present (Figs. 4 and 6, A and D). Simulated mu rhythm exhibited current flow that oscillates up and down the cortical layers symmetrically around zero (Fig. 6, C and E). Alternating excitatory synaptic drive to dendrites in the infragranular layers (FF) and supragranular (FB) layers drives subthreshold intracellular current flow up and down the PN dendrites.
(3) Mu-alpha and mu-beta components often occurred at different times (Figs. 4 and 5). Simulated mu rhythm contained mu-alpha and mu-beta components that had the same simultaneity characteristics as those of the MEG data (Figs. 6C and 7B). Stochastic in timing of each FF and FB “input burst” (Fig. 2), arriving about every 100 ms, changes the relative dominance of the net FF and FB input strength on each cycle. Enhancing the FF inputs increases the mu-alpha power, whereas enhancing the FB inputs increases the mu-beta power (Fig. 8).
(4) As predicted initially by the model, tactile evoked responses during high prestimulus mu-, compared with low-mu conditions, exhibited an early M50 positive peak, followed by a trend toward a decreased M70 peak (Fig. 9B). Simulated SI evoked responses via a sequence of FF (∼25 ms post-stim) followed by FB (∼70 ms) and LFF input (∼135 ms) input during high mu, compared with low mu (Fig. 8E) exhibited an early M50 positive peak and a decreased M70 peak (Fig. 9A). High-mu states created an ongoing depolarization in the PNs and INs, causing greater firing in both populations to the initial FF input. Back-propagation of action potentials in the PNs created the M50 peak. Recruited inhibition in turn decreased the subsequent excitatory cell response to the approximately 70-ms FB input, decreasing the M70 peak (Fig. 9C).