H minimal changes in the overall SR3029 biological activity calcium dynamics. While eliminating the oscillation without affecting release is unfeasible in the 
laboratory, the protocol we developed allows us to implement it in the mathematical myocyte model via a dynamic clamping of the variables involved. We describe first the details of the dynamic clamping of the SR calcium load and then of the level of recovered RyR2s. Both clamping protocols can be activated separately or simultaneously. In the latter 
case, cytosolic calcium alternans should disappear if there is no other intervening mechanism.Clamping of the SR Ca LoadFigure 1 illustrates the workings of this protocol. Initially, the cell is paced at a given rate using our standard numerical model until calcium alternans has stabilized (see Figure 1A). Then, starting at any given beat, we change the Homatropine methobromide dynamics of the sarco/ endoplasmic reticulum Ca2+-ATPase (SERCA) uptake during the last 150 ms of diastole so that the SR calcium load achieves the same value c?SRClamp before each external excitation (see Figure 1B). The value of c?SRClamp is taken to be equal to the maximum presystolic SR calcium load during the non-clamped dynamics. During the external excitation and SR calcium release, the original SERCA current is used. In this way we do not affect the dynamics of calcium release and re-uptake. More importantly, the dynamics is only affected when all the variables are close to their equilibrium values. As seen in Figure 1B the result is a clamped dynamics where the SR Ca load before each calcium release is constant and where the evolution of the cytosolic calcium transient (Figure 1C) indicates if calcium alternans is affected or not by this clamping. The procedure can be summarized as testing whether calcium alternans disappears when the SERCA current is set as follows: JSERCA Q10{SERCA Vmaxi KmfHc 1z K imfSR H { Kmr H SR H z Kmr??fornTvtv(nz1)T{t0 msJSERCA 10Vmax ( ?SRClamp { SR ) for (nz1)T{t0 vtv(nz1)T ms??Dynamic Clamping ProtocolsDuring alternans, the intracellular cytosolic calcium transient alternates from beat to beat. Whenever this happens there is a corresponding alternation in both the pre-systolic SR calcium load and the level of RyR2 ready to open (not inactivated, i.e. in state R of Figure S1 in Appendix S1). These two types of oscillations are directly related with two mechanisms proposed to account for calcium alternans in the literature. One states that a change in the calcium loading process leads to cytosolic calcium alternans (calcium alternans due to SR calcium load). The other states that the level of RyR2s recovered from inactivation oscillates. An ideal experiment to discern the underlying mechanism would require eliminating the alternation in one of thewhere typically t0 = 150 ms (we use t0 = 75?00 ms for T,240 ms). Equation (1) is the original SERCA uptake, with the parameters given in [17] active during the external excitation, calcium release and first stages of the reuptake. During the last t0 before each beat we substitute the SERCA uptake for a stronger current, which keeps pumping calcium from the cytosol until SR ?SRClamp .Clamping of RyR2 RecoveryFollowing the same idea of the previous clamping protocol, the clamping of RyR2 recovery is achieved changing its dynamics during the 150 ms before each calcium release (see Figure 2). These changes in the RyR2 are applied to eliminate dynamically the oscillation in the pre-systolic ratio of recovered RyR2 (R state)Ca2+ A.H minimal changes in the overall calcium dynamics. While eliminating the oscillation without affecting release is unfeasible in the laboratory, the protocol we developed allows us to implement it in the mathematical myocyte model via a dynamic clamping of the variables involved. We describe first the details of the dynamic clamping of the SR calcium load and then of the level of recovered RyR2s. Both clamping protocols can be activated separately or simultaneously. In the latter case, cytosolic calcium alternans should disappear if there is no other intervening mechanism.Clamping of the SR Ca LoadFigure 1 illustrates the workings of this protocol. Initially, the cell is paced at a given rate using our standard numerical model until calcium alternans has stabilized (see Figure 1A). Then, starting at any given beat, we change the dynamics of the sarco/ endoplasmic reticulum Ca2+-ATPase (SERCA) uptake during the last 150 ms of diastole so that the SR calcium load achieves the same value c?SRClamp before each external excitation (see Figure 1B). The value of c?SRClamp is taken to be equal to the maximum presystolic SR calcium load during the non-clamped dynamics. During the external excitation and SR calcium release, the original SERCA current is used. In this way we do not affect the dynamics of calcium release and re-uptake. More importantly, the dynamics is only affected when all the variables are close to their equilibrium values. As seen in Figure 1B the result is a clamped dynamics where the SR Ca load before each calcium release is constant and where the evolution of the cytosolic calcium transient (Figure 1C) indicates if calcium alternans is affected or not by this clamping. The procedure can be summarized as testing whether calcium alternans disappears when the SERCA current is set as follows: JSERCA Q10{SERCA Vmaxi KmfHc 1z K imfSR H { Kmr H SR H z Kmr??fornTvtv(nz1)T{t0 msJSERCA 10Vmax ( ?SRClamp { SR ) for (nz1)T{t0 vtv(nz1)T ms??Dynamic Clamping ProtocolsDuring alternans, the intracellular cytosolic calcium transient alternates from beat to beat. Whenever this happens there is a corresponding alternation in both the pre-systolic SR calcium load and the level of RyR2 ready to open (not inactivated, i.e. in state R of Figure S1 in Appendix S1). These two types of oscillations are directly related with two mechanisms proposed to account for calcium alternans in the literature. One states that a change in the calcium loading process leads to cytosolic calcium alternans (calcium alternans due to SR calcium load). The other states that the level of RyR2s recovered from inactivation oscillates. An ideal experiment to discern the underlying mechanism would require eliminating the alternation in one of thewhere typically t0 = 150 ms (we use t0 = 75?00 ms for T,240 ms). Equation (1) is the original SERCA uptake, with the parameters given in [17] active during the external excitation, calcium release and first stages of the reuptake. During the last t0 before each beat we substitute the SERCA uptake for a stronger current, which keeps pumping calcium from the cytosol until SR ?SRClamp .Clamping of RyR2 RecoveryFollowing the same idea of the previous clamping protocol, the clamping of RyR2 recovery is achieved changing its dynamics during the 150 ms before each calcium release (see Figure 2). These changes in the RyR2 are applied to eliminate dynamically the oscillation in the pre-systolic ratio of recovered RyR2 (R state)Ca2+ A.