Articles in PresS. J Neurophysiol (October 1, 2014). doi:10.1152/jn.00914.2013
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45
Generation of field potentials and modulation of their dynamics through volume integration of cortical activity Yoshinao Kajikawa1 & Charles E. Schroeder1,2 1. Cognitive Neuroscience and Schizophrenia Program, Nathan S. Kline Institute for Psychiatric Research, Orangeburg, NY. 2. Department of Psychiatry, Columbia University College of Physicians and Surgeons, New York, NY Running Head: Generation
and modulation of cortical field potentials
Corresponding author: Yoshinao Kajikawa Nathan Kline Institute 140 Old Orangeburg Rd Orangeburg, NY10965 Phones: +1 (845) 398-6630 Fax: +1 (845) 398-6545 E-mail:
[email protected] Abstract: Field potentials (FPs) recorded within the brain, often called “local field potentials” (LFPs), are useful measures of net synaptic activity in a neuronal ensemble. However, due to volume conduction, FPs spread beyond regions of underlying synaptic activity, and thus, an “LFP” signal may not accurately reflect the temporal patterns of synaptic activity in the immediately surrounding neuron population. To better understand the physiological processes reflected in FPs, we explored the relationship between the FP and its membrane current generators using current source density (CSD) analysis in conjunction with a volume conductor model. The model provides a quantitative description of the spatiotemporal summation of immediate local and more distant membrane currents to produce the FP. By applying the model to FPs in the macaque auditory cortex, we investigate a critical issue that has broad implications for FP research. We show that FP responses in particular cortical layers are differentially susceptible to activity in other layers. Activity in the supragranular layers has the strongest contribution to FPs in other cortical layers, and infragranular FPs were most susceptible to contributions from other layers. In order to define the physiological processes generating FPs recorded in loci of relatively weak synaptic activity, stronger synaptic events in the vicinity have to be taken into account. While outlining limitations and caveats inherent to FP measurements, our results also suggest specific peak and frequency band components of FPs can be related to activity in specific cortical layers. These results may help improving the interpretability of FPs. Keywords: Field potential, Current source density, Volume conduction, LFP.
Copyright © 2014 by the American Physiological Society.
46
Abstract
47 48
Field potentials (FPs) recorded within the brain, often called “local field potentials” (LFPs), are
49
useful measures of net synaptic activity in a neuronal ensemble. However, due to volume
50
conduction, FPs spread beyond regions of underlying synaptic activity, and thus, an “LFP” signal
51
may not accurately reflect the temporal patterns of synaptic activity in the immediately
52
surrounding neuron population. To better understand the physiological processes reflected in
53
FPs, we explored the relationship between the FP and its membrane current generators using
54
current source density (CSD) analysis in conjunction with a volume conductor model. The model
55
provides a quantitative description of the spatiotemporal summation of immediate local and more
56
distant membrane currents to produce the FP. By applying the model to FPs in the macaque
57
auditory cortex, we investigate a critical issue that has broad implications for FP research. We
58
show that FP responses in particular cortical layers are differentially susceptible to activity in
59
other layers. Activity in the supragranular layers has the strongest contribution to FPs in other
60
cortical layers, and infragranular FPs were most susceptible to contributions from other layers.
61
In order to define the physiological processes generating FPs recorded in loci of relatively weak
62
synaptic activity, stronger synaptic events in the vicinity have to be taken into account. While
63
outlining limitations and caveats inherent to FP measurements, our results also suggest specific
64
peak and frequency band components of FPs can be related to activity in specific cortical layers.
65
These results may help improving the interpretability of FPs.
66
67
Introduction
68
The field potential (FP) recorded within active neural tissue is an information-rich measure that
69
arises mainly from synaptically-driven transmembrane currents related to excitability
70
fluctuations in ensembles of neurons (Buzsaki et al. 2012; Mitzdorf 1985). Decoding of FPs can
71
extract information about motor commands (Bansal et al. 2012; Ince et al. 2010), sensory stimuli
72
(Belitski et al. 2010; Kayser et al. 2007) and current behavioral and cognitive state (Fries et al.
73
2001; Lakatos et al. 2009, 2013; Scherberger et al. 2005; Steriade et al. 1993). This is true
74
whether the FP is tightly or loosely phase-locked to an identified stimulus [so-called “evoked”
75
and “induced” FPs (Makeig et al. 2004; Shah et al. 2004)], and it also holds for “spontaneous”
76
activity, whose immediate antecedents are obscure (Fukushima et al. 2012; Lakatos et al. 2005).
77
Despite its strengths, the interpretation of FPs is complicated by volume conduction, the
78
spread of electric field generated by current sources in a conductive medium. As FPs are
79
typically recorded with an active recording electrode amplified against a reference signal from a
80
distant electrode, the exact point of origin of the signal is unknown due to volume conduction
81
(Einevoll et al. 2013; Nunez and Srinivasan 2006). The signal could in theory arise anywhere
82
within the conductive medium housing the electrodes. Understanding of the principles of volume
83
conduction allows one to track down the brain sources of electroencephalographic (EEG) and
84
event related potential (ERP) components (Godlove et al. 2011; Schroeder et al. 1995). This
85
same understanding also prompts the use of techniques that isolate relatively local activity like
86
current source density (CSD) (Einevoll et al. 2013; Kajikawa and Schroeder 2011; Mitzdorf
87
1985; Nicholson and Freeman 1975, Pettersen et al. 2006) or the surface Laplacian, a 2-
88
dimensional approximation of the second derivative of potential, used to derive scalp current
89
density (Nunez and Srinivasan 2006; Tenke and Kayser 2012). CSD analysis estimates the
90
laminar profile and magnitude of current sources (net local outward currents) and sinks (net local
91
inward currents) that produce fluctuation of voltage in the conductive medium of the
92
extracellular space (Nicholson and Freeman 1975), and is in most experimental cases explored
93
by analyzing the one-dimensional spatial pattern of FPs (but see Riera et al. 2012). Previous
94
ERP/EEG studies have shown that ERP peak components and EEG frequency components
95
(bands) have distinct scalp topographies and source localizations. The same conditions apply to
96
the laminar-spatial patterns of FP and CSD signals within a cortical area. This is a critical
97
problem for LFP studies, as it limits their ability to attribute experimental effects to specific cell
98
populations (e.g., supragranular pyramidal cells) or input types (e.g., top-down vs. bottom-up)
99
Effects of volume conduction at a macro (>1 mm) scale are recognized as essential to the
100
formation of the spatiotemporal profile of electromagnetic signals that can be measured in ERP
101
and EEG recordings at the scalp (Nunez 1998; Nunez and Srinivasan 2006; Pascual-Marqui
102
1999), but these effects at a micro (< 1 mm) scale are poorly understood (Bédard and Destexhe
103
2011, Buzsaki et al. 2012, Einevoll et al. 2013, Riera et al. 2012). Recent papers on the related
104
topic of the “spatial spread of the LFP” have outlined several variables that likely affect volume
105
conduction beyond the boundaries of active (generator) tissue. Some have emphasized spatial
106
variables, such as the size and shape of the generator substrate that is the population of neurons
107
whose transmembrane currents are reflected in an FP, and the magnitude of the activation
108
(Kajikawa and Schroeder 2011), while others have emphasized temporal variables, such as the
109
synchrony of local cellular activity (Lindén et al. 2011; Reimann et al. 2013). Critically, in most
110
circumstances, synchrony and strength of activation are confounded. On one hand, a given
111
ensemble of neurons is expected to produce a stronger net response, as reflected by a larger FP
112
signal magnitude, when they are activated simultaneously than when they are activated
113
asynchronously (e.g. Reimann et al. 2013). On the other hand, we cannot directly relate the
114
magnitude of FP signal, to the degree of synchronous firing in a given neuron population
115
because, as discussed above, synchrony is not a sole factor that determines spatial spread of the
116
FP. In the present study, we deal with the spread of FP activity within hot spots of highly
117
synchronized activation; i.e., we examine local neuronal responses to preferred stimuli that
118
excite massive and synchronous activation of local neuron ensembles. In this circumstance, high
119
synchrony can be assumed. Interestingly, such strong, synchronous, responses, likely
120
encompassing multiple cortical layers, are reported in many papers. However, investigation of
121
how local (i.e., sub-millimeter scale), well-correlated (i.e. sensory-evoked response) activity
122
patterns of neuron populations in different layers of cerebral cortex, contribute to the
123
construction of a local FP began only recently.
124
A simulation study investigated the contributions of electrical activity of neurons and
125
their compartments to LFP at such scales and suggested the potential role of the active
126
membrane conductance that generates action potentials in shaping the spatiotemporal patterns of
127
FP (Reimann et al. 2013). Regardless of types of membrane conductance and difference in
128
patterns of constituent neurons, transmembrane currents of neuronal population can be translated
129
to CSD features. In sensory systems, sensory stimuli effective in driving strong neuronal firing
130
responses also evoked larger CSD responses (Fishman and Steinschneider 2006, Kajikawa and
131
Schroeder 2011) and larger postsynaptic currents (Tan et al. 2004), that corroborate the close
132
correspondence between CSD components and transmembrane currents.
133
Detailed understanding of the contribution of local activity to local FP is particularly
134
important at this time, as studies are beginning to investigate laminar and sublaminar
135
distributions of FPs in different frequency bands in order to help to sort out influences of
136
feedforward and feedback projections (Buffalo et al. 2011; Spaak et al 2012; Sundberg et al.
137
2012). Importantly in this regard, distinct frequency bands can be dissociated by their differential
138
behavior in both event-related responses (Fries et al. 2011; Khayat et al. 2010; Ray and Maunsell
139
2011; Wilke et al. 2006) and ongoing activity (Lakatos et al. 2005; Magri et al. 2012). However,
140
the spatiotemporal summation by which these microprocesses combine to form a local FP has
141
not been directly investigated in physiological data. In an effort to advance the understanding
142
and physiological interpretation of FP measures this study examined the contribution of volume
143
conduction to the generation of FPs at a sub-millimeter scale. We used a volume conductor
144
model constructed on physical principles to spatially integrate “observed” CSD activity
145
distributed across cortical depths, in order to calculate a “predicted” spatiotemporal pattern of
146
intracortical FPs, and evaluated the similarity between the originally observed and model-derived
147
FP patterns. Our results indicate that volume conduction integrates distributed activity across
148
layers to shape the temporal pattern of an FP in any given layer. Our results also suggest that
149
particularly in the infragranular layers where local FP generation is relatively weak, the form and
150
frequency content of the locally-recorded FP may largely reflect activity in other cortical layers.
151 152
Materials and Methods
153 154
All experimental procedures were approved by the Institutional Animal Care and Use Committee
155
of the Nathan Kline Institute.
156
Subjects. Six macaque monkeys were implanted with a headpost and one or two recording
157
chambers during aseptic surgeries. The chambers were oriented to make penetrations
158
perpendicular to the lateral sulcus.
159
Stimuli. Calibrated binaural tone stimuli were delivered through two free field speakers directed
160
to ears (TDT). Tones ranged from 353.55 Hz to 32 kHz with 0.5 octave intervals (14
161
frequencies) were presented at 60 dB SPL (duration: 100 ms, SOA: 625 ms).
162
Recordings. All recordings were done while monkeys were awake. We used electrodes of linear-
163
arrays of 23 electrical contacts spaced either 100 or 200 m apart (0.3-0.5M at 1.0 kHz) to
164
record signals at different cortical depths. FP (0.1Hz-500Hz) were recorded from all contacts
165
simultaneously and sampled at 2 kHz. As a reference (or indifferent) electrode, a metal pin
166
immersed in saline filling the recording chamber was used. The best frequency (BF) of multiunit
167
activity responses to tones was identified in each recording site (Kajikawa and Schroeder 2011).
168
In the present study, responses to BF tones were analyzed.
169
Current source density. CSD was calculated from FPs recorded from 3 adjacent electrode
170
contacts by the second order finite differences:
171
CSDz FPz dz 2FPz FPz dz dz 2 ,
(Eq. 1)
172
as the one dimensional approximation of the second order spatial derivative of the FP (Mitzdorf
173
1985), with an assumption of homogeneous conductivity (see Discussion):
174
2 2 2 2 r 2r 2 2 2 y z z 2 x
,
(Eq. 2)
175
The generality of this method of CSD approximation (Eq. 1) has been questioned recently
176
because of generator “edge” effects in rodent barrel cortex (Einevoll et al. 2013), with regard to
177
traditional assumption of “trans-positional invariance” to validate Eq. 1 performed only in
178
vertical dimension across cortex (Mitzdorf 1985). However, the solution proposed by Einevoll
179
and colleagues “iCSD,” (discussed below) was devised for the unique structure of rodent barrel
180
cortex, and does not generalize to the neocortex as a whole. Here, to support the last equality in
181
Eq. (2) as an approximation, “trans-positional invariance” is only a special case of generally
V V V , z x y
182
required conditions that satisfy the inequality:
183
electrodes (see Discussion). As Figure 2D will show, the condition is usually met when
184
suprathreshold stimulation excites extensive regions of cortex (see also Reimann et al. 2013).
185
Thereby, it is appropriate for a wide range of neocortical applications, including the present case.
186
In the present study we used this standard approximation for compatibility with the vast majority
187
of prior studies.
188
Spectral analyses of FP and CSD. To derive the spectrotemporal patterns and frequency bands of
189
those FP and CSD ( FP(t ) and CSD(t ) respectively), we used the complex Morlet wavelet
190 191
t ' 4
1
e
i0t '
, along the vertical path of array
t' 2
e , 0 6 ,
to calculate the wavelet transforms of those signals (Torrence and Compo 1998):
192
WTLFP (t , f )
193
WT CSD(t , f )
194
f
t
FP(t 't ) s
t
*
t 't dt ' , s
CSD(t 't ) s
*
t 't dt ' , s
(Eq. 3),
0 2 02 , 4s
195
in which t is the sampling interval. The complex Morlet wavelets were scaled to conserve the
196
energy relationship between frequency components. The translation step was 1 ms. Center
197
frequencies of the wavelet stepped from 1 to 256 Hz with 0.1 octave intervals. This dense
198
frequency sampling leaves high redundancy between neighboring frequency bands of the
199
wavelet transform and allows reconstruction of the original signal as (Farge 1992)
200
256 FP(t ) Re WTFP ( f , t )df . 1
201
Power and phase spectrotemporal distributions were derived as the absolute values and phase
202
angles of WTFP (t , f ) and WTCSD (t , f ) .
203
Frequency band signals of FP and CSD were calculated as the sums of 16 center
204
frequencies (delta: 1.0~2.9 Hz, theta: 3.0~8.8 Hz, gamma1: 27.7~81 Hz, gamma2: 83.9~256 Hz)
205
or 8 center frequencies (alpha: 9.1~15.9 Hz, beta: 17.0~25.8 Hz,) of WTFP (t , f ) and WTCSD (t , f )
206
along the frequency axis (Kajikawa and Schroeder 2011). Derived signals were approximately
207
1.6 or 0.8 octaves wide with small overlaps between frequency bands, and were complex-valued,
208
same as WT (t , f ) . The real part of signals (e.g. WT (t, f ) ) was identical to the transform with
209
real part of wavelet (e.g. replacing t ' in Eq. 2 with t ' ). As the original signals were of
210
real-valued, real parts of derived signals corresponded to the filtered signals and were used as
211
frequency band-limited signals (Farge 1992). Absolute amplitudes of the complex transforms
212
were used as the power of signals.
213
Volume conductor model. The volume conductor model assumed that CSD at different depths
214
contributed to the FP by volume conduction. Under the assumption of homogeneous tissue
215
conductivity, volume conducted signals decay in proportion to the inverse of distance with no
216
assumption of frequency dependent gain change; this predicted decay is matched by empirical
217
findings (Schroeder et al. 1992; 1998; Kajikawa and Schroeder 2011). We calculated FP at a
218
certain depth d k that could be generated by a given spatial distribution of CSD at every moment
219
as
220
vcFP k , t A j
CSDd j , t h d j dk 2
2
, h rh 2d d
(Eq. 4),
221
in which d d k 1 d k is the spatial distance between neighboring contacts of array electrode.
222
The parameter h represents a displacement distance of the center of mass of CSD from the array
223
electrode assuming all CSD components were vertically aligned across cortical depths. The idea
224
behind the displacement is as follows.
225
In general, there are many possible solutions of sources that could generate topographic
226
or depth patterns of FP in solving Eq. (2). In the present case, the only boundary was the line of
227
the array electrodes with depth-distributed signals with surrounding open space. It was not
228
definitive in finding particular CSD solutions as there are many possible configurations of CSD.
229
However, to be consistent with the columnar structure of cortex, one constraint we could
230
presume was that CSD distributed vertically across cortical layers. That reduced the problem to
231
one of determination of the finite horizontal spread. However, again, any horizontal spread of
232
CSD could be replaced by a variety of other CSD to create the same spatial pattern of FP. Thus,
233
it would be an ideal approach to reduce the arbitrariness of horizontal spread and model it as
234
simply as possible, since at this time, we cannot realize more complex (parametric) models in
235
detail or justify them.
236
Responses to suprathreshold pure tones in auditory cortex must distribute as medial-
237
lateral stripes defined by iso-frequency contours. The shapes of stripes were kept similar across
238
cortical layers and horizontally finite, constrained by the anterior-posterior extent of an iso-
239
frequency domain. Any such spatial distributions of sources may be substituted by one or several
240
charge densities located at a horizontal point somewhere inside of responsive zones to generate
241
same field potential patterns. That is equivalent to what is done at a different spatial scale in
242
source estimation for EEG/MEG. The anchoring point around which spread is estimated is single
243
point at the center of mass in a horizontal distribution of sources. Its exact position may remain
244
unknown. Only its distance from electrodes is needed in deriving FP at the electrodes. Thus, we
245
included the horizontal displacement parameter h to represent the horizontally delimited
246
distribution of activity in the simplest way. While such a simplified model can introduce error
247
(uncertainty), it is preferable to a model with more unknown parameters, unless they can be
248
empirically validated.
249
The output of the resultant model becomes equivalent to that generated by a dipole-like
250
substrate (i.e., having a vertical distribution), horizontally displaced at an average distance h
251
from the recording electrode array across depths (Kajikawa and Schroeder 2011; Somogyvari et
252
al. 2012). In Eq. 3, rh 2 d was the sole free parameter that could influence the shape of vcFP (k , t ) .
253
Its values were chosen to maximize the similarity (see below) of spatiotemporal patterns between
254
the calculated and the observed LFP. The parameter h rh 2d d may be considered to be the
255
distance to the center of horizontally distributed sources generating FP from the electrodes. Note
256
that, due to the linearity of both wavelet decomposition and volume conduction, the order of
257
calculating them is interchangeable. However, once the power or phase of wavelet-decomposed
258
signals is derived, these values cannot be entered into the volume conductor model due to the
259
non-linearity of power/phase calculations.
260
Within a “column,” vertical profiles of CSD that contributed to the FP at a given depth
261
were derived as follows. The spatiotemporal CSD profiles were weighted along the depth axis by
262
their distances from the origins of CSD with an additional displacement h . Wavelet-derived
263
frequency band signals were derived in a similar manner. The spatiotemporal profiles of CSD
264
power were derived after distance-weighing.
265
Depth distributions of the power of CSD contributions to granular FP responses and
266
phase differences of CSD from granular FP responses were derived as follows. The mean of the
267
CSD power over a150 ms period post-onset of tone was calculated for each CSD and subtracted
268
by mean power during a 150 ms period prior to the tone onset. For data sampled with 200 m
269
intervals, missing data points were linearly interpolated. For each frequency band, such
270
distribution was normalized by the power at the depth of the granular sink. Since depths of
271
cortical layers relative to electrode positions differed between penetration sites, depths of
272
granular sinks were aligned to zero. Median and confidence intervals at each depth were derived
273
after alignment.
274
While the field generated by dipoles decays with the inverse of the square of distances, 2
275
1 r , the volume conductor model treated individual poles of multi-poles as separate monopoles
276
whose field decayed with the inverse of the distance, 1 r , not squared. However, it should be
277
noted that superimposition of electric fields generated by two monopoles of opposite signs
278
decaying with 1 r becomes same as an electric field that decays with 1 r
279
relatively larger than the dimension of the dipole. Therefore, while the model treats CSD features
280
as spatially distributed monopoles, it can generate the field potentials generated by dipoles as
281
well.
282
This implementation was preferred, because recordings in the present study were done within the
283
zones of active generators or dipoles, where the dipole field approximation’s relevance is
284
questionable.
285
Similarity score. To quantify the similarity between temporal patterns of FP and CSD, and the
286
similarity between the spatiotemporal patterns of the observed FP and CSD or model-derived FP,
287
we calculated the similarity score as described below. Temporal patterns of FP, CSD, and model-
288
derived FP at individual recording channels were expressed as vectors, vFP , vCSD and v vcFP of nT
289
independent samples, where nT = sample duration x sampling rate. Spatiotemporal profiles were
290
expressed as matrices, M FP , M CSD , and M vcFP , whose number of rows corresponded to the
2
at distances
291
number of recording contacts of electrodes - 23 and 21 for FP and CSD, respectively.
292
Spectrotemporal patterns of signals at each channel were expressed similarly in matrices
293
WTFP , WTCSD , and WTvcFP ; whose number of rows corresponded to the number of wavelet
294
center frequencies. Before calculation of the similarity, these vectors and matrices were
295
normalized by the root of mean squared values (e.g. v
296
compensate for their difference in the signal power and to analyze shapes in isolation. For
297
matrices, the mean was calculated across both rows and columns. Similarity between those
298
patterns was calculated as the inner product between vectors or the Frobenius inner products of
299
matrices after normalization. The similarity value ranged from –1 for completely mirrored
300
shapes of opposite polarities to 1 for identical shapes of same polarities regardless of
301
magnitudes. While similarity could depend on the periods of FP, CSD, or their wavelet
302
transforms, values were derived using signals between -30 and 170 ms from the onset of sound.
303
However, the similarity score is model-free in that it can be evaluated for any models or analysis
304
that creates spatiotemporal patterns as long as it keeps dimensions (the number of depth channels
305
and time bins) constant.
306
Statistics. All statistical tests were done non-parametrically with a criterion level of p = 0.01,
307
except for ad hoc comparisons at p = 0.05. 95% Confidence intervals were derived by bootstrap
308
using the Matlab function, bootstrp, with 1000 re-sampling (Figs. 4, 6A and B).
2 mean v or M
) to
mean M
2
309 310
Results
311 312
Sensory-evoked FP and CSD signals at a given location in the brain often differ in temporal
313
pattern even though both signals are considered to reflect local synaptic activity. We first
314
elaborate the difference between these signals in both temporal and spectral domains. We then
315
describe how the volume conductor model bridges the gap between the two signals. We further
316
show that the FP integrates local and distant activity in individual frequency bands independently
317
of the spatial distributions of the CSD activity of other frequency bands.
318 319
Differences between co-located FP and CSD signals.
320
First, we show the difference in temporal patterns between FP and CSD. Figure 1 shows
321
examples of FP and co-located CSD responses recorded simultaneously from (i) supragranular
322
(Sg), (ii) granular (Gr), and (iii) infragranular (Ig) layers in an A1 penetration, along with
323
spectrotemporal patterns of signal power (C,D) and phase difference (E). FPs in the different
324
layers share similar patterns starting with onset negativity (Fig. 1A), but after this point
325
waveforms diverge between Sg and other layers. This is not surprising because FP patterns are
326
expected to change systematically as recordings traverse an active FP generator region (Mitzdorf
327
1985, Schroeder et al. 1995) and the strongest CSD activity was found in the Sg layer. The point
328
of interest here, however, is the contrast between co-located FP and CSD signals as it reveals
329
effects of volume conduction. The FP - CSD contrast is described systematically by comparison
330
of their temporal patterns (B), low frequency power plots (C vs. D) and phase differences (E). In
331
Sg layers, both FP and CSD signals maintained negative deflections after the onset until 100 ms.
332
Low frequency power had similar peaks. In Gr layers, though temporal patterns looked similar,
333
the onset negativity occurred earlier in CSD than FP and had a nearly 90 degrees phase
334
difference in high frequency bands. In Ig layer, the larger positive/negative fluctuations in FP
335
were more rapid than the source/sink fluctuations in the CSD signal. Thus, the peak FP power
336
was higher in a higher frequency band than peak CSD power.
337
To assess the dissimilarity between FP and CSD across penetration sites, we quantified
338
the similarity score as an inner product of vectors representing temporal patterns of 2 signals in 3
339
cortical layers (Fig. 1E). The score changes from 1 when FP and CSD had identical temporal
340
patterns to -1 when two patterns were of opposite polarity and mirror images of one another. The
341
scores for 3 pairs in Fig. 1B were 0.85 (i), 0.44 (ii), and -0.08 (iii), respectively. The distributions
342
of the score were skewed towards 1 in all cortical layers. Median scores were 0.81, 0.75, and
343
0.33 for Sg, Gr, and Ig layers respectively (n=130), with significant difference (Friedman’s non-
344
parametric repeated measures ANOVA, 2(df=2, n=116) = 30.4, p < 0.01, excluding penetration
345
sites missing Ig layer). This difference was attributable to the lower scores of Ig layer when
346
compared to other layers (Tukey's HSD test, p < 0.05). Thus, temporal patterns of FP and CSD were
347
more dissimilar in Ig layers than Gr and Sg layers. However, even in Gr and Sg layers, the scores
348
distributed over wide ranges. These results suggested the generality of the differences in
349
spectrotemporal patterns between co-located FP and CSD signals. Below, we show how the
350
discrepancies between the two signals are explained by volume conduction, and evaluated the
351
effect by the similarity score.
352 353 354
Volume conductor model Here, we explain the volume conductor model first, and then fill the gap between FP and
355
CSD by applying the model to experimental data. CSD signals reflect transmembrane current
356
flow patterns that are localized to the spatial extent of neuronal processes (1). The local 1-
357
dimensional CSD profile is customarily derived from a laminar FP profile using Eq. 1 by
358
approximating 2 (r ) qr , in which is the tissue conductivity (assumed to be spatially
359
uniform), (r ) is the FP, and qr is CSD. FP reflects spatially distributed events qr through
360
volume conduction. When the boundary condition of () 0 is met, 2 (r ) qr can be
361
solved as 1 qr r ( r ) dr . 4 r r
362
(Eq. 4)
363
Since auditory cortical FP responses to best frequency tones are several hundreds of microvolts
364
and scalp ERPs centimeters away from the auditory cortex are generally on the order of 1
365
microvolts, we used this solution as an approximate model of a local volume conductor. There is
366
no constraint on the spatial distribution of qr within a body. The volume conductor model
367
summates spatially distributed CSD signals after weighing them with the inverse of their
368
distances from the point of measurement as in Eq. (4), which weakens distant CSD to some
369
degree.
370
Figure 2 shows how Eq. (4) governs volume conduction when there are two different FP
371
generators, each consisting of sink and source, operating in isolation (one strong source/sink -
372
Fig. 2A; one weak sink/source - Fig. 2B), or together (Fig, 2C). Sink sand sources modeled here
373
have round shapes delimited to two parallel flat planes spread horizontally (xy, Fig. 2Ai and Bi).
374
FP profiles were calculated based on Eq. 4. Due to symmetry of sink and source shapes around
375
their centers, FP profiles formed in planes that contain x-y centers of sink and source are
376
identical. When there is one FP generator, inversion of FP polarity occurs at the mid z position
377
between sink and source (Fig. 2Aii and Bii). Above and below the inversion, the FP spreads with
378
constant polarity with amplitude peak positions aligned vertically to sink and source (Fig. 2Aiii
379
and Biii). However, when two FP generators co-exist, FP distributions from single generators
380
superimpose, and an FP distribution reflecting an interaction of the two generators emerges.
381
Ventral spread of the negative FP below the inversion of the single strong generator (Fig. 2A) is
382
reduced in the combined case (Fig. 2C) and is interrupted by small positivity due to the lower
383
generator’s source. Compared to the positive peak below the inversion in Fig. 2B, the positive
384
peak at the corresponding z-position is smaller in Fig. 2C. Thus, within the cortex where activity
385
spreads over cortical layers through volume conduction, the FP in any one layer is
386
‘contaminated’ by signals of activity generated in other layers. We utilized the volume conductor
387
model to specify this effect quantitatively.
388
Figure 3 shows how the volume conductor model links sensory-evoked FP and CSD
389
responses. At all penetration sites, we simultaneously recorded FP at 23 depths spanning the
390
superficial to deep layers of auditory cortex, as illustrated by the representative FP profile in Fig.
391
3C, from which the spatiotemporal CSD profile in Fig. 3A was derived. The volume conductor
392
model-derived FP profile for these signals (described below) is shown in Fig. 3B. Fig. 3D
393
illustrates the gradual change in the FP produced by the systematically increasing spatial range of
394
the CSD profile that is incorporated into the volume conductor model to derive Gr layer FP
395
waveform. This is shown in increasing order from top to bottom, starting from a summation of
396
one channel that corresponds to Gr layer CSD itself. While temporal patterns of the observed FP
397
and CSD (bottom red and top green traces in Fig. 3D) differed in their temporal patterns, the FP
398
generated by the model (traces between the top and bottom in Fig. 3D) became successively
399
closer approximations of the observed FP as the spatial range of CSD estimates (incorporated by
400
the model) is increased. In this example, the model broadened the onset negativity and delayed
401
the timing of the following positive peak (asterisks in Fig. 3D) from 38 to 62 ms. This effect
402
indicates that the Gr layer FP was influenced by current generators (indexed by CSD) in other
403
layers. By performing the same model calculation over all depths of FP recording, we calculated
404
a model-derived spatiotemporal FP profile (Fig. 3B), which resembles the profile of the observed
405
FP (Fig. 3C) more closely than it does the CSD profile. Accordingly, the spectrotemporal pattern
406
of the model-derived FP for the Gr layer also bears greater similarity to that of the observed FP
407
than that of the CSD signal (Figs. 3E-G).
408
For the example shown in Fig. 3, the similarity score of the observed FP with the co-
409
located CSD signal was 0.38 and increased to 0.95 when computed using the model-derived FP.
410
Also, across penetration sites (n=130), the similarity over the 150 ms long temporal patterns
411
between the observed FP in Gr layer and the model-derived FP increased as a function of the
412
number of proximal CSD signals (Fig. 4B). Improvement of the score through use of the model
413
waveform was also observed for FP in Sg and Ig layers (Fig. 4A and C).
414
In all layers, the score reached asymptote by integrating signals within less than 1 mm.
415
This suggested that FP responses to the best frequency tones within the auditory cortex are
416
shaped by activity of variable temporal patterns occurring in the gray matter over the range of a
417
millimeter where strong activation is spread. It may be also noted that the model-derived FP in
418
Sg layer, where local CSD activity was stronger than other layers (Fig. 6C), still increased its
419
similarity to the observed FP by integrating activity over several hundreds of micrometers. These
420
results suggested that even at loci where the CSD signals were strong, the effect of volume
421
conduction would not be negligible due to activity in their vicinity.
422
Notably, the FPs in Ig layers (which are farthest from Sg layers) had the lowest similarity
423
scores. FPs in these layers required the widening of spatial coverage in the model to include CSD
424
in the Sg layers in order for the similarity score to asymptote. We estimated the channel
425
increments needed for the scores to reach 90% of the net changes, for all 3 layers in each
426
penetration site (Fig 4D). Those channel increments differed significantly between layers
427
(Friedman’s nonparametric repeated-measures ANOVA, 2(df=2, n=116) = 38.2, p < 0.01,
428
excluding sites missing Ig layer data). This difference was attributable to the channel increments
429
needed significantly more in Ig layer than other layers (Tukey's HSD test, p < 0.05). Thus, in
430
order to define the physiological processes generating FPs recorded in loci of relatively weak
431
synaptic activity (e.g., the Ig layers), stronger synaptic events in the vicinity (e.g., those in the Sg
432
layers) have to be taken into account.
433 434 435
Contributions of different frequency bands to volume conducted signals Below, we expand the consideration of volume conduction to the relationships between
436
components of FP and CSD. Sensory cortical responses are not only distributed across layers but
437
also transient and of broad spectral content. A CSD signal qr , t can be considered as a linear
438
summation of the CSD frequency bands q f r , t as q r , t q f r , t . By incorporating this into
f
439
440
Eq. 4, FP signals decomposed into frequency bands in the volume conductor model is 1 (r , t ) 4
q r r , t f
f
r r
dr .
441
Due to linearity of both volume conduction for spatially distributed CSD signals and spectral
442
transforms for frequency bands in each CSD signal, those operations are interchangeable. Thus,
443
the FP due to volume conduction can be derived separately for different frequency bands as
444
1 q f r r , t f (r , t ) dr . 4 r r
445
Figure 5 illustrates how spatiotemporal patterns of CSD activity in the different
446
frequency bands contribute to the infragranular FP response. Fig. 5A and B show the
447
spatiotemporal profiles of CSD responses and their power over a wide frequency range (,
448
1~256 Hz) and in five 1.6 octave- (, , and ) and 0.8 octave-wide ( and ) frequency
449
bands. Due to division by the distance from the point of interest, the volume conductor model
450
weakens distant CSD before summation. Therefore, actual signals in the Sg layer CSD
451
components were stronger than what is plotted in Fig. 5A and B, regardless of frequency band.
452
In each frequency band, summation of CSD responses weighted inversely by the distance from
453
the Ig layer along depth reflects the Ig layer FP response derived by the volume conductor model
454
(orange traces in Fig. 5C).
455
In the example shown in Fig. 5, the Ig layer CSD response (green in Fig. 5C, ) was
456
negative from 10 to 100 ms after the onset of sound and had a negative peak at 50 ms followed
457
by a positive peak at 130 ms. The FP response (red in Fig. 5C, ) had a positive peak at 50 ms
458
followed by a negative peak around 130 ms. Temporal pattern differences between CSD and FP
459
signals were reflected in opposite polarities essentially throughout the response time course
460
across frequency bands (, , and ). Beta band CSD and FP signals also differed in
461
frequency.
462
These temporal pattern discrepancies between CSD and FP signals across frequency
463
bands were due to volume conduction. In most frequency bands, powers of signals in Gr and Sg
464
layers were comparable to or even stronger than the power in the Ig layer (Fig. 5B, , , , and
465
). In each band, by spatially summating the CSD signals across depths, the volume conductor
466
model reproduced an FP response that greatly resembled the observed FP, but with a polarity
467
opposite to that of the co-located Ig layer CSD signal (orange in Fig. 5C). Thus, the temporal
468
pattern of the Ig layer FP response reflected not the local generator currents, but rather those in
469
other cortical layers via volume conduction.
470
We quantified the relative contributions of CSD signals across depths to the granular FP
471
for all experiments (penetration sites). In each penetration site, the depth distributions of CSD
472
response power, weighted by the inverse of distances from the Gr layer sink were calculated
473
separately for the same 6 frequency bands as done for Fig. 5. In general, sensory evoked CSD
474
signals are distributed across cortical layers in sensory cortices and have broad frequency content
475
(Kajikawa and Schroeder 2011; Lakatos et al. 2005, Maier et al. 2011).
476
Figure 6A illustrates the laminar distribution of CSD power by frequency band that
477
contributed to the Gr layer FP response. As the power of the Gr layer CSD responses differed
478
significantly between frequency bands (median values from low to high frequency bands: 0.49,
479
0.41, 0.076, 0.032, 0.012 and 0.86x10-3 mV/mm2, Friedman's test, 2(df=5, n=130) = 621, p