Beta Null Spikes, Phase Cones, and Vortices

 

Walter J Freeman

http://sulcus.berkeley.edu/Null_Spikes_Apices/

 

ElectroCorticoGraphic fields of potential at the surface of the visual cortex, recorded with a 6x6 mm electrode array with spacing of 0.79 mm from a rabbit trained to discriminate visual conditioned stimuli

 

00. Summary description of DVD cinematographic representations of ECoG

(00SummaryDescription,DVD,ECoG,BetaPhaseTransition.doc)

 

There are three pairs of DVDs, each giving a perspective view and a contour plot of the three main state variables: the amplitude of the filtered ECoG in that pass band, the analytic amplitude, and the analytic phase.

 

01. Summary of Time Functions: to be Viewed Concomitantly with Spatial Displays

 (01F152X12test20.602to21.8sec20to25HzPhaseTransition.jpg)

 

A. PSD by the multitaper method showing peaks in theta but not beta or gamma ranges.

Long time segments (³6 sec) give PSD with power-law distributions, which reflects the scale-free dynamics of cerebral cortex (Freeman and Breakspear, 2007). The optimal pass band is 5 Hz, as shown by the symbol Ô – Ô at the frequency band selected for analysis in the mid-beta range.

B. The superimposed 62 ECoG filtered signals show maximal spatial coherence in two high-amplitude bursts of oscillation separated by minima referred to as beats in Rayleigh noise. The null spike at the minimum in the envelope of the oscillations 21.098 sec is an example of a phase transition. Not all Rayleigh beats manifest phase transitions. The line segment shows the time range of the movies.

C.  Log10 analytic power shows a succession of down spikes, the deepest of which are referred to as null spikes (Freeman, 2009). The latencies from the arbitrary start of the display are in sec

D. The analytic frequency in rad/sec is approximated by the ratio of the change in phase, DF in radians, in each time step to the duration of the digitizing step, Dt, in sec. The spikes reflect indeterminacy when the analytic amplitude is low; the flat segments between spikes show times of minimal temporal and spatial variance of the analytic phase. The length of the flat segment determines the duration of a frame.

 

A0. Perspective, Electric Field of a Burst Demarcated by Null Spikes

(A0F152X12test21.0to21.6sec20to25HzFilteredECoGFlapping.mov)

 

The movie shows a perspective image of the scalar electric potential field in the mid-beta 20-25 Hz range from 21 sec to 21.6 sec (0.6 sec), with the amplitude color-coded. The task for analysis is to identify the stable components of phase and analytic power (analytic amplitude squared) of multiple bursts of oscillations at differing center frequencies and mean phase. The tendency for rotation prevents use of PCA and ICA, which assume time-invariance of the signals over space, i.e., spatial stationarity.

 

A1. Electric Potential Field: Complex Rotating Amplitude Patterns (Vortices)

(A1F152X12test21.0to21.sec20to25HzRotationECoG,CCW,Flat.mov)

 

The movie shows the view from above with amplitude minus the spatial average, color-coded to show the rotation of the field in complex patterns. The widespread coordination is strong evidence that the neural field underlying the scalar fields of potential is a spatiotemporal continuous vector field that influences and includes every neuron in the region of phase-lagged synchrony in the cortex. Every point has both amplitude and vectorial directions of temporal and spatial change.

 

B0. Examples are Shown of Null Spikes, often in Clusters

(B0F152X12test21.0to21.6sec20to25HzLog10AnalyticPowerPersp.mov)

 

This movie shows a perspective representation of the log10 analytic power from 21 sec to 21.6 sec (0.6 sec) across the three clusters of down spikes.  The color-coding shows the fluctuating amplitude and phase of the filtered signal (not the analytic amplitude or power). The temporal and spatial localization of the down spikes is apparent. 

 

B1. Contour Plots Show the Locations of Null Spikes (Blue)

(B1F152X12test21to21.6sec20to25HzPhaseExplosionCCW.avi)

 

The same data as in B0F are shown as seen from vertically over the surface and color-coded to represent power: blue is minimum. The spatial localization is limited by the interelectrode distances averaging 0.79 mm in an approximately 6x6 mm array placed on the surface of the rabbit visual cortex.

 

C0. A Phase Cone Appears Shortly after a Null Spike in the Analytic Phase
(C0F152X12test21.0to21.8sec20to25HzPhasePersp.mov)

 

The phase gradient is radially symmetric about a center point that closely conforms to the location of the preceding null spike. The maximum lead of the analytic phase designates the apex of a cone with downward slope, indicating a phase lag to the periphery called an explosion. The approximate co-location of the null spike and the apex of the following phase cone (either maximal lead = explosion, or maximal lag = implosion) is seen only in frames in which a single burst has one dominant frequency.

 

C2. The Conic Apex is near the Location of the Null Spike and a Phase Discontinuity

(C1F152X12test21to21.6sec20to25HzPhaseExplosionCW.avi)

 

The same data as in C0F are presented as seen from vertically above the cortical surface, color-coded with maximum +¹ and minimum –¹ radians. The features of note are the evidence for phase discontinuity in the temporal vicinity of the null spike near 21.1 sec and the array numbers ordinate 6 and abscissa 5.

 

References

 

Freeman W.J. [2004a] Origin, structure, and role of background EEG activity. Part 1. Analytic amplitude. Clin. Neurophysiol. 115: 2077-2088.

      http://repositories.cdlib.org/postprints/1006

 

Freeman W.J. [2004b] Origin, structure, and role of background EEG activity. Part 2. Analytic phase. Clin. Neurophysiol. 115: 2089-2107.

http://repositories.cdlib.org/postprints/1486.

 

Freeman WJ [2007] Hilbert transform for brain waves. Entry for Encyclopedia for Computational Neuroscience, Izhikevich E (ed.)

      http://www.scholarpedia.org/article/Hilbert_transform_for brain_waves

 

Freeman WJ, Breakspear M [2007] Scale-free neocortical dynamics. Encyclopedia for Computational Neuroscience, Izhikevich E (ed.).

     http://www.scholarpedia.org/article/Scale-free_neocortical_dynamics

Freeman WJ (in press) Evidence for a spatiotemporal singularity in percept formation by cerebral cortex. Proc ICCN2009, 15-19 Nov, Hangzhou, China.   ICCN2009,WJFreeman Singularity