Supplemental data

Files in this Data Supplement:

• supplemental material - Supplemental Fig. S1 Site dependence in resting membrane potential in layer 5 neurons. Comparison of the dependence of resting membrane potential on distance from the soma in the model neuron (filled circles) and experimental data (open circles). Both experimental (black line) and model (dashed line) data could be well fitted with a linear regression with a slope of 1.4 mV/100 �m. In the absence of Ih the resting membrane potential in the model was uniform (grey dots).
• supplemental material - Supplemental Fig. S2 Effect of external potassium on Ih currents and single-channel properties. A, Ih currents recorded in dendritic cell-attached patches 850 �m from the soma with the indicated K+ and Na+ concentrations in the patch-pipette (in mM). Reducing external K+ from 120 mM to 60 mM diminished the maximum conductance by approximately one third (from 145 pS/�m2 to 52 pS/�m2). Under physiological extracellular K+ concentrations (2.5 mM K+/120 mM Na+) the average Ih conductance density was further reduced to 29 pS/�m2. B, Example of a variance-mean plot of Ih recorded with 60 mM K+ external. Same data as shown in A (60/60). Parameters derived from NSFA are indicated. C, Pooled data of the effect of external K+ on Ih current density (closed symbols) and channel number (open symbols). NSFA with 120 and 60 mM K+ indicated that only N decreased significantly from ~1209 to 577 (P < 0.01, 500-850 �m from soma). The single-channel conductance was not affected (674 � 35 vs. 723 � 210, P > 0.9). Furthermore, consistent with a lack of effect of K+ on Ih activation (Hestrin, 1987; Hille, 2001) Po was unchanged (0.99 to 0.90 � 0.03, P > 0.4). Ih amplitudes under physiological ionic conditions were too small to reliably estimate the single-channel parameters using NSFA (see Supplemental Figure 1). Based on these data, under physiological conditions we measured a 5.1-fold reduction in Ih conductance density compared to conditions with 120 mM external K+, primarily due to a reduction in available channel number.
• supplemental material - Supplemental Fig. S3 Statistical reliability of variance-mean parameter estimation. To investigate the influence of a large range of different signal-to-noise conditions on the reliability of single-channel parameters extracted using NSFA we analyzed simulated Ih currents using our non-fluctuating model with varying numbers of channels (N = 10 � 2000) and single-channel amplitudes (i = 5 fA � 1 pA). 1 pA r.m.s noise at 10 kHz frequency bandwidth was added giving single-channel signal-to-noise ratios of 0.005 to 1. Data were digitally filtered at 100 Hz. The plots can be summarized in the following way: If the amplitude of the macroscopic Ih current is ≤ 1 pA or i ≤ 10 fA, the errors in i, N, and Po were larger than 50% and thus highly inaccurate. However, with mean Ih current amplitudes > 10 pA, as typically found in this study, the errors are ≤ 20%. We conclude NSFA with a fc = 100 Hz has adequate precision to estimate the single-channel parameters of Ih channels under our recording conditions. Error (%) = 100 | (xfit-xsim) | / xsim, where xfit is the value estimated by NSFA and xsim is the actual value used in the simulation. The black bars represent errors > 50%.