The experimental results discussed here (most of it) concern data from two models: the isolated chick retina and the in situ rat hippocampus. In the vertebrate retina, the Intrinsic Optical Signal (IOS) associated with retinal spreading depression waves (RSDs) is macroscopic [Note 1], making this preparation unique for the observation of its propagation. The most probable patterns are solitary circular waves and logarithm spirals. Circular spirals were rarely seen and stay mostly in place with a little jitter, whereas the logarithm spirals wander around the retina creating another level of abstract patterns (see [11] ). Circular, two dimensional propagation can be taken mathematically as a special case of a dynamic spiral pattern propagation.

Figure 1 and Figure 2 show the birth and propagation of a solitary circular retinal wave. At the inset the figure shows two time series that were recorded simultaneously with the video in the Picture In Picture (PIP) mode. One series records the extracellular potential at the electrode tip, the second the mean bright brightness of a square pixel matrix with 50 µm side length overlaying the electrode tip. This time series shows the optical changes or the Intrinsic Optical Signal (IOS). Note that the field potential leads the IOS in this wave (shown in Figure 2). This pattern is frequent; however, also frequent is the simultaneous rise of both signals.

Ten seconds later the wave is a large circle of optical changes, what one sees is the loss of retina transparency and consequent increase in light scatter. Figure 2 shows how the extracellular drop of potential (by historical convention in electrophysiology, negativity is up and due to autoscaling of the recording system, one sees the time series moving in the downward direction) relates to the optical changes in tissue around the electrode tip. Remembering from the previous figure, the wave propagates at the rate of 45 µm/sec. It exactly at this distance from the electrode tip that the wave changes the extracellular potential. One second later that dV/dt will peak (frames 127 - 129). It is at this time that the dIOS/dt catches up and also peaks.

Figure 1 and Figure 2 show a circle of increase in light scatter; however, the IOS associated with membrane excitation have light scatter and birefringence variations [9] [12] tightly coupled to ionic activity changes.

The temporal evolution of ionic activity associated with wave propagationresult in time series in most experiments. The following relationship exists among them: the peak of the dV/dt = d[k]/dt = d[Na]/dt whereas d[Ca]/dt begins at the peak of dV/dt (see for example [3] [4] [5] [12] [13] ).

The calcium activity variation is tightly coupled to a component of the pH changes. This component happens at the frontwave and has a kinetics of seconds and thus cannot be associated with metabolism. It must be something else. Our interpretation is that protons and hydronium ions “disappearance” is the result of dissociated water becoming liquid water. Therefore, the pH shift at the front wave, or at least the major part of it.

The dissociated water makes the Exclusion Zone (EZ) at membranes; this state of water is denser than the liquid and has birefringence [14]. Both calcium activity and pH changes were recorded in RSDs as changes in fluorescence images [2] [4] [5] [12]. The novelty here was that wave patches of tissue could lack the ionic change response maintaining the light scatter at 545 nm. A second slow pH change associated with RSDs is metabolic-production and secretion of lactic acid by microglia, the kinetics follow the second optical component of the optical profile of RSDs [2] [5] the time scale is minutes. In summary: the ionic activity changes in the frontwave are all compatible with the proposed charged gel (EZ) changes and that includes the early alkalinization that lacks explanation in the HH membrane model of same data.

The kinetics of potassium activity and potential drop are tightly coupled in spreading depression waves; to the point that many researchers attributed entirely to potassium channels activity the potential drop (see for example [13] ). Not only that, the potential and ion activity changes have the same signal at both sides of the macroglial membranes. The same is true for the early pH shift, the same alkalinization happens at both sides of glial cells membranes.

The latter variables early pH and potassium activity variables are also tightly coupled to the optical properties changes in chick retinas. In chick retinas in particular a structural feature is of importance: about 2/3 of the length of the Müller cells are divided in fine tubes [5] these tubes have no organelles like mitochondria but are filled with intermediate filaments made of GFAP (Glial Fibrillary Acidic Protein) see also [14]. The telltale part is the acidic mention. In here we assume that around the stack of the acid fibrils, water dissociates and creates and EZ (exclusion zone) [15]. Among the properties of EZs is birefringence. At the quiescence/excited state transition this EZ relaxes and in consequence, the Müller cells lose their lightguide role transiently due to birefringence loss [12] [16] [17] and light scatter increases creating the optical changes or IOS.

In other words, the frontwave releases energy. It does so in several forms of flows, in which water flows are an important part both liquid and dissociated water (for example hydronium ions), free cations and semi mobile anions (radicals in membrane carbohydrates and proteins). The macroscopic circular current that arises at the transition to propagating wave releases red/near infrared light besides heat [8] and when propagation begins green light scatter appears (505 - 540 nm) [5] [12]. We observed this transition many times with voltage sensitive dyes and looking at the Intrinsic Optical Signal (IOS) with infrared and black/white cameras associated with optical filters (see Fig. 9 to Fig. 11 of [5] ). We discovered that the red/near infrared scatter followed the potential distribution, therefore a circular linear pattern implies a circular current. When waves were elicited by a light mechanical touch, standing patterns of IOS emerged around the stimulated area and preceded the circular current (see Fig. 17 and Fig. 19 of [2] ).

The purpose of Figure 3 and Figure 4 is to show how the transition can be followed by observing the tissue IOS.

Figure 3 (left frame) shows a mechanical stimulation of central retina using a tungsten needle (50 µm). Pressing a charged (piezoelectric) gel will move charges, therefore, mechanical and electrical stimulation are very similar. The mechanical one has an exact point of stimulation, showed in Figure 3. Close to three seconds later, a full blow, perfectly circular propagating wave is present around the stimulation point. The diameter of this wave is 300 µm. Figure 4(a) to Figure 4(f) follow the transition. In the stimulated region a standing dynamical pattern arises hundreds of milliseconds after the touch. The light scatter associated with it contains only red/near-infrared wavelengths. The dynamic character of the spatial pattern conveyed in shimmering bright points inside it and in lower frequency space/time oscillations of brightness of the whole pattern. When a circle closes, Figure 4(d), the IOS includes the blue/green part of the light scatter appears (peak at 545 nm) and propagation begins, Figure 4(e). Because the red/near-infrared part of the IOS matches the field potential distribution, one can see that a circular current closes and initiates propagation. What was lacking was the semblance of the magnetic field associated with the circular current. To realize that a magnetic torus guided water and ions flow, we had first to make experiments with liquid deuterium [18] [19] and interpret these results in the light of the energy release at the wavefront. Second, we looked for experimental models that studied water release of energy in the form of electromagnetic flows, heat and mechanical flow of water. This model exists and is the subject of study of Elmar C. Fuchs and cols. [20] [Note 3].

First let us examine the liquid deuterium results. We applied liquid deuterium as the solvent in both isolated retinas and Belouzov-Zabotinsky reaction system. As was the case with similar comparative experiments [21] the response of both systems was very similar, both responded with collapse of excitability in a short time (2 hours). The pH of liquid water at 25 degrees Celsius is 7.0, the liquid deuterium pH is 7.4 or deuterium dissociation is five times smaller than that of liquid water dissociation. In our interpretation of the transition from quiescence to propagating wave, these results were not surprising as we first though, but the expected effect of the solvent physical properties changes. It should be noted that the classical HH membrane model does not either predict or explain the solvent deuterium effects, by contrast, Ichiji Tasaki model does [9].

Second, if we examine the flows in a water bridge, at the center, hydronium and hydroxyl ions run toward the cathode and anode electrodes, however twisting around the central current there is mass flow of liquid water and added particles configuring a torus [20]. In water bridges and CNS wavefronts, these charge and mass flows are coupled.

In summary, mechanical stimuli induce a dynamic standing pattern in which fluctuations in release of energy are amplified until a circular current form within the tissue, this current follows the electrical field. In parallel, a magnetic flow guides mass water and particles flow in the shape of a torus. This magnetic torus we propose as a key factor in spreading depression wave propagation.

We also reasoned that time series associated spatial propagation patterns in other parts of the CNS, had to be compatible with such patterns if our conjecture was correct, (see [22] ). It was fortunate that in one study of propagation of spreading depression and after discharges, the sampling rate was 700 Hz, very unusual in ECoG recordings [10]. The geometry of the macroglial network in rat hippocampus is cylindrical and that is important. Here we assume that our interpretation about the role of the geometry of the macroglial network in the shape of spreading depression waves is correct [22]. Cylindrical geometry (think of windsocks in small airports) implies Von Karmann Street vortices, they in turn assumed to optimize the energy expenditure within the cylinder. If one measures this pattern in several points in the form of time series, then a split in frequencies should be present. Reexamining the data from the ECoG study [10] at the transition from after discharge to spreading depression, we found the pattern compatible with the expected frequency split of cylindrical geometry vortices. Our conjectures stand.

A further validation can look at vortices presence in excitable tissue and other electrochemical systems such as the B-Z system. Yes, vortices were found in heart tissue and B-Z [23] [24]. Besides these results, another set of experiments suggests vortex effects, non-stationary pathways in B-Z [25] and in isolated retinas [26]. In retinas, the sudden (step) change in potassium activity in the maintenance solutions from 6 to 1or 0 mEq/l, produces a hyper excitable pathway with a succession of spreading depression waves that terminate in tissue death in about 30 minutes. These waves have altered IOSs. Last, but not least, the alteration of the isolated retinal solution by addition of the liquid glycerol adds viscosity and polarity to the retinal bath solution. The spread velocity, field potential and the IOS observed in the presence of glycerol differed from controls. The red/near infrared light scatter was present whereas the green component was depressed, the spread velocity was slower and voltage record changed kinetics with a prolonged plateau (see [4] [5] [12] ). Once more, the electric equivalent model and molecular biology do not predict or explain these results. By contrast, interference with flows do.

Lashley [6] used the scotomas (bright columns) march through his visual field to predict spreading depression waves as we said in the introduction section. Therefore, the knowledge that the frontwave distorts perception is as old as the wave discovery. Also it is long known that these wave have cognitive consequences in the form of learning and memory mechanisms blockage (for review see Bures, J.1974 [27] Sparks and Gallo, M. 2007 [28] ); Bures and Buresova, 1963 [29]; Travis, R.P. and Sparks, D.L. 1963 [30] ).

These cognitive impairments occur with structural reversible changes in the gels associated with membranes and energy expenditure in the form of heat and red/near infrared non-Plank dissipation, mass and charge flow. One question arises: is this the maximum rate of dissipation possible? Several sets of data about cortical and retinal waves appear to support that, indeed, this is the case. In the first 10 - 12 seconds of a wave one cannot distinguish between a reversible excitation wave and a non-reversible pathway to tissue death; see, for example [31]; later on, the pattern of spread and the irreversible character of an excitotoxic response makes them clearly distinct. The initial state of several pathways appears the same and thus, vortex effects will explain them all: the magnetic flow guides the mass (carried by liquid water) and charge flows.

Another question emerges: can resonances in the form of eddies of similar flows or Von Karmann street explain a perceptual phenomenon like the epiphanies?

Epiphanies with visualization of problems solutions exist in science. Famous are the ones related by August Kekule and the benzene ring and what Roger Penrose told about visualizing quasi-crystals before chemistry was aware of them [32]. What characterizes these perceptual states are the strong positive emotion (elation) and a strong memory trace let for recall, in contrast with the memory loss associated with the state at the frontwave of spreading depression waves. Here we speculate that an unusual resonant state could be present in the brain during the few seconds long insight period. The spatial shape of the resonances varying in different parts of the brain. The tubular spinal chord and brainstem candidates for Von Karman Street pattern.

Vortex effects appear to be the nature way of maximizing energy expenditure; their expression present in weather phenomena like tropical storms and hurricanes. Storms can stay put or drift slowly, whereas hurricanes do travel, albeit with different velocities depending on many factors, among which the temperature gradient. The same mechanisms for energy expenditure are present in both phenomena. In the cylindrical rat hipocampus, we found the predicted frequencies split at the frontwaves of spreading depressions [10] [33], either this is a happy coincidence or vortex effects are indeed present in the brain.