34). Two physiological factors likely can account for these differences.
First, due to reflection of subthreshold synaptic currents in the CSD measure, the tuning of CSD responses to tones is wider than that of MUA responses to the same tones. Second, due to volume conduction of electrical events in auditory cortical loci tonotopically not matched to the penetration sites, the tuning of LFP is wider than that of CSD measures. The idea that LFP responses to tones octaves away from the BF at a penetration site in A1 is due to volume conduction predicts that the CSD index derived by numeric differentiation from such an LFP profile should not contain the “volume conducted” components. In other words, the local spatiotemporal distribution of sources and sinks outlined by CSD analysis would not be able to generate the observed profile of VX770 LFP response (LFPobs). To test this idea, laminar LFP responses
(LFPcal) were calculated back from CSD profiles. According to Poisson’s differential equation, see more the local LFP profile is the spatial integration of its solution given a particular spatial distribution of current sinks/sources identified by CSD analysis (Experimental Procedures). Figure 4A (left column) shows laminar-temporal profiles of LFPobs responses to tones, in a penetration site tuned toward low frequencies. The profiles maintained common patterns across tone frequencies: the predominant onset negativity in the bottom two-thirds of channels and positivity in the top one-third of channels across tone frequencies. Other later features, like the strong positivity around 50 ms in the bottom of the profile, were preserved only for responses to lower frequency tones. CSD responses (Figure 4A, second column) are similar until to LFPobs in terms of their strength across frequencies below 1.4 kHz. However, CSD responses are nearly abolished at high stimulus frequencies. Tuning curves in Figure 4B also show that CSD responses were nearly zero at high stimulus frequencies where LFPobs responses still had amplitudes
about 20% of peak values. Figure 4A (third column) shows laminar-temporal profiles of LFPcal derived from CSD profiles using Equation 1 (Experimental Procedures). Note that our simultaneous recording from single arrays orthogonal to cortical layers cannot resolve the fine details of spatial distributions for sinks/sources. For example, lateral spread of activity may differ between layers, but cannot be elucidated by our methods. Regardless, application of Equation 1 to CSD worked qualitatively well to calculate LFP when that was generated locally. LFPcal at low frequencies had largely similar profiles to LFPobs from the onset to the later inversions of polarity across similar subsets of the recording depths. As the tone frequency increased, the response became weaker.