Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2015 Aug 4;9:37.
doi: 10.3389/fncir.2015.00037. eCollection 2015.

Periodotopy in the Gerbil Inferior Colliculus: Local Clustering Rather Than a Gradient Map

Affiliations
Free PMC article

Periodotopy in the Gerbil Inferior Colliculus: Local Clustering Rather Than a Gradient Map

Jan W H Schnupp et al. Front Neural Circuits. .
Free PMC article

Abstract

Periodicities in sound waveforms are widespread, and shape important perceptual attributes of sound including rhythm and pitch. Previous studies have indicated that, in the inferior colliculus (IC), a key processing stage in the auditory midbrain, neurons tuned to different periodicities might be arranged along a periodotopic axis which runs approximately orthogonal to the tonotopic axis. Here we map out the topography of frequency and periodicity tuning in the IC of gerbils in unprecedented detail, using pure tones and different periodic sounds, including click trains, sinusoidally amplitude modulated (SAM) noise and iterated rippled noise. We found that while the tonotopic map exhibited a clear and highly reproducible gradient across all animals, periodotopic maps varied greatly across different types of periodic sound and from animal to animal. Furthermore, periodotopic gradients typically explained only about 10% of the variance in modulation tuning between recording sites. However, there was a strong local clustering of periodicity tuning at a spatial scale of ca. 0.5 mm, which also differed from animal to animal.

Keywords: auditory midbrain; functional anatomy; inferior colliculus; periodic sound; periodotopy; pitch; tonotopy.

Figures

Figure 1
Figure 1
Multi-unit responses quantified using either the “analog multi-unit activity (MUA)” method, or by counting threshold crossings with time bins as indicated. Responses to click trains presented at the click rates indicated to the left are shown. Each trace is the average response over 40 trials.
Figure 2
Figure 2
Waveforms (left) and cochleagrams (right) of the three types of periodic stimuli used in this study. The cochleagrams were calculated by passing the stimulus waveforms through a gammatone filter bank, and provide an estimate of the basilar membrane (BM) deflection expected along the tonotopic array (y-axis) as a function of time (x-axis).
Figure 3
Figure 3
(A) Schematic diagram of the electrode array used in this study. (B) Example of a typical multiunit frequency response area (FRA) recorded in ICc. The blue-to-yellow color scale indicates the multiunit response strength observed for pure tones presented at the frequency and intensity indicated on the x- and y-axis respectively. The best frequency (BF) indicated by the green arrow was measured by summing responses for stimulus amplitudes >50 dB SPL. (C) FRAs collected during one medio-lateral multi-electrode array penetration through the inferior colliculus (IC). Each of the 64 squares is one FRA, with x- and y-axis ranges as in (B). Channel 27 is highlighted by a green frame, as it is the same data as that shown in more detail in (B). FRAs from recording sites outside the ICc are shown in a desaturated color map. (D) BFs derived from the data shown in (C) plotted as a color-coded tonotopic map superimposed to scale, onto a Nissl stained histological section through the corresponding part of the IC. The color of the filled circles indicates the BF for each recording site according to the color scale on the right. Large filled circles show BFs of recording sites located within ICc, small circles indicate recording sites outside ICc. The axes show coordinates in mm from the mid-line and the dorsal edge of the midbrain respectively. DC, dorsal cortex of the IC; LS, lateral shell of the IC; PAG, periaqueducal gray. (E) FRAs as in (C), recorded during a multielectrode penetration oriented along the sagittal plane. Two of the channels on this multielectrode array were faulty, hence two FRAs are missing, one at the top right, the other in the bottom row. (F) Composite showing the BFs obtained from the data shown in (E), as well as the data from a second recording location, 0.5 mm deeper, superimposed to scale onto a photo-micrograph of a sagital section through the gerbil IC taken from page 21 of Cant and Benson (; by kind permission of the publisher). SCs, superior colliculus, superficial layers; SCd, superior colliculus, deep layers.
Figure 4
Figure 4
(A) Tonotopy in ICc. Each 8 × 8 line grid shows the positioning of one 64 channel array electrode placement, relative to the top, medial and rostral edge of the ICc. Each colored dot shows an electrode recording site inside the ICc. The color of the dot indicates the BF recorded at the corresponding location. A systematic progression from low BFs (blue colors) near the top to high BFs (red colors) near the bottom is readily apparent. These data were subjected to linear regression analysis to determine the best fit tonotopic axis of the gerbil ICc. (B) Multiunit BF plotted against location along the tonotopic axis. Data from each of the six animals in our study were plotted using a different symbol and color. A random Gaussian distributed jitter with a standard deviation of 0.09 octaves has been added to the BF values to spread out overlapping data points.
Figure 5
Figure 5
(A) Examples of responses from a typical multiunit with a pure tone BF of 7.49 kHz to click trains (left, blue), SAMN (middle, green) and IRN (right, red) for the full range of fundamental frequencies tested. Each line shows the median response amplitude over 25 presentations of the stimulus with modulation frequency (MF) as shown on the left. The x-coordinate represents post-stimulus time, with the black bars at the bottom indicating the duration of stimulus presentation (200 ms). (B) Rate modulation transfer functions (rMTFs) derived from the data shown in (A) by plotting the total response (area under the curves in A) as a function of stimulus MF. Error bars show standard error of the mean. Standard errors were small, so error bars may be hard to see. It is evident that both the shape and the maxima of the rMTF were strongly stimulus dependent; the rBMFs for this multiunit were 2510 kHz for click trains and IRNs, but only 95 Hz for SAMN. (C) Temporal modulation transfer functions (tMTFs), calculated as described in the “Materials and Methods” Section, for the responses shown in (A).
Figure 6
Figure 6
Rate tuning to periodicity in click trains (blue), SAMN (green), and IRN (red) for two sample electrode penetrations, one oriented medio-laterally (A) the other oriented rostro-caudally (B). Each small square shows the responses for the corresponding recording site in the 8 × 8 electrode array. Each colored curve gives a tuning curve, with stimulus MF on the x-axis and response amplitude on the y-axis. Axis labels were omitted as they would be too small or too crowded to see, but the scaling is identical to that in Figure 5B, i.e., the abscissas cover a range of [−5,3] octaves re 500 Hz, and the ordinates cover the range [−0.1 × y_max, y_max], where y_max is the maximum response observed across all stimulus types and recording sites in the corresponding multielectrode penetration. The curves are shown with error bars (mean + standard error) but error bars are very small and may be hard to see. Tuning curves from recording sites inside the ICc are plotted against a white background, those outside the ICc are shown against a light gray background.
Figure 7
Figure 7
(A) Best modulation frequencies shown as 3D color maps (top row) and as scatter plots of response best modulation frequency (rBMF; the “best” modulation frequency which evoked the maximal response) against distance along the best fit periodotopic axis (bottom row). Data are shown for click train (left), SAMN (middle) and IRN (right) stimuli. The 3D maps are laid out as in Figure 4A, with the axes showing the anatomical coordinates of the recording sites. Each black grid shows the position of one multi-electrode array penetration. The colored dots indicate the location of multiunits that were located within the ICc, and for which changes in stimulus MF accounted for at least 10% of the response variance. The color indicates the MF of the stimulus that evoked the strongest response (compare color scale on the right). (B) Scatter plots laid out as in Figure 4B, with different colored symbols used to plot the data recorded from each animal. A random Gaussian distributed jitter with a standard deviation of 0.09 octaves has been added to the rBMF values to spread out overlapping data points. The number of multiunits in each data set (N) and correlation coefficient (R) are shown above each panel. Triple asterisks (*) indicate that the R values are statistically significant at p < 0.001. (C) Periodotopic maps constructed from multiunits within a functionally defined iso-frequency lamina (BFs within 15% of 1 kHz). The ordinate and abscissa give anatomical distances in mm within the iso-frequency plane. The black dots show the positions of multiunits that were excluded from the analysis because their responses were not strongly influenced by stimulus MF. The white dots show the location of multiunits whose rBMFs were used to construct the periodotopic maps by 2-D interpolation.
Figure 8
Figure 8
Scatter plots comparing BF (x-axis) against rBMF obtained for each multiunit with each of the three types of periodic stimuli (as indicated on the axes). Data from different animals are plotted using dots of different colors, and a random Gaussian distributed jitter with a standard deviation of 0.09 octaves has been added to the rBMF values to spread out overlapping data points. The number of multiunits in each data set (N) and correlation coefficient (R) are shown above each panel.
Figure 9
Figure 9
Scatter plots comparing the rBMFs obtained for each multiunit with each of the three types of periodic stimuli (as indicated on the axes). As in Figure 8, data from different animals are plotted using dots of different colors to show that the distribution of rBMFs was comparable across animals; a random Gaussian distributed jitter with a standard deviation of 0.09 octaves has been added to the rBMF values to spread out overlapping data points. The number of multiunits in each data set (N) and correlation coefficient (R) are again shown above each panel.
Figure 10
Figure 10
(A) Histograms plotting the distribution of correlation coefficients between pairs of periodicity tuning curves, where pairs were chosen either randomly (“random pairs”, light blue markers) or from recording sites with anatomical coordinates that were no more than 0.3 mm apart (“near pairs”). The two tuning curves in the “near” pairs came from nearby coordinates either in a single animal (“within animal”, green lines) or from two different animals (“across animals”, red curves). (B) 2-D histogram illustrating the dependence of the similarity (correlation) between periodicity tuning curves on anatomical distance. In the heat maps, each row shows the distribution of correlation coefficients for pairs of tuning curves recorded at points separated by an anatomical distance within 1/16 of a mm from the bin center shown on the y-axis. For small anatomical distances, large positive correlation coefficients predominate (i.e., tuning properties are very similar for the large majority of nearby pairs recording sites). This positive skew declines with increasing anatomical distance. The light blue dots show the mean for each histogram. Their error bars cover a range of ±3.3 SEM. Given the large N, some of the error bars may be too small to see. The light blue vertical lines demarcate the overall mean correlation coefficient across all sample pairs at all distances.
Figure 11
Figure 11
Phase-locking to the stimulus period for click trains (blue), SAMN (green), and IRN (red) for the same two sample electrode penetrations for which rate tuning curves were shown in Figure 6, one oriented medio-laterally (A), the other oriented rostro-caudally (B). Each colored curve shows the phase-locking coefficient as a function of stimulus MF. Tuning curves from recording sites inside the ICc are plotted against a white background, those outside the ICc are shown against a light gray background. Axis labels were omitted as they would be too small or too crowded to see, but all axes are scaled identically to Figure 5C, i.e., the abscissas cover a range of [−4,1] octaves re 500 Hz and the ordinates cover a range of [0, 1].
Figure 12
Figure 12
Phase-locking curves for all multiunit recordings in our sample, in response to click-trains (left), SAMN (center) or IRN (right), sorted by low-pass cutoff frequency. Each row shows the phase-lock coefficient according to the color scale on the right as a function of stimulus MF (x-axis) for one multiunit. Low-pass cutoff frequencies were defined as the stimulus MF for which the phase-lock coefficient dips below a value of 0.4, and these are shown as black dots for each phase-lock tuning curve.
Figure 13
Figure 13
Anatomical distribution of phase-lock limits (low-pass cutoff MFs) shown as 3D color maps (top row) and as scatter plots of cutoff MF against distance along a best fit anatomical axis (bottom row). The layout of the figure is analogous to that of Figure 7. Triple asterisks (*) indicate that the R values are statistically significant at p < 0.001.
Figure 14
Figure 14
(A) Histograms plotting the distribution of correlation coefficients between pairs of tMTFs, where pairs were chosen either randomly (“random pairs”, light blue markers) or from recording sites with anatomical coordinates that were no more than 0.3 mm apart (“near pairs”). The two tuning curves in the “near” pairs came from nearby coordinates either in a single animal (“within animal”, green lines) or from two different animals (“across animals”, red curves). (B) 2-D histogram illustrating the dependence of the similarity (correlation) between tMTFs on anatomical distance. In the heat maps, each row shows the distribution of correlation coefficients for pairs of tuning curves recorded at points separated by an anatomical distance within 1/16 of a mm from the bin center shown on the y-axis. The light blue dots show the mean for each histogram. Their error bars cover a range of ±3.3 SEM. Given the large N, some of the error bars may be too small to see. The light blue vertical lines demarcate the overall mean correlation coefficient across all sample pairs at all distances.

Similar articles

See all similar articles

Cited by 12 articles

See all "Cited by" articles

References

    1. Baumann S., Griffiths T. D., Sun L., Petkov C. I., Thiele A., Rees A. (2011). Orthogonal representation of sound dimensions in the primate midbrain. Nat. Neurosci. 14, 423–425. 10.1038/nn.2771 - DOI - PMC - PubMed
    1. Bizley J. K., Walker K. M. M., Silverman B. W., King A. J., Schnupp J. W. H. (2009). Interdependent encoding of pitch, timbre and spatial location in auditory cortex. J. Neurosci. 29, 2064–2075. 10.1523/JNEUROSCI.4755-08.2009 - DOI - PMC - PubMed
    1. Brand A., Urban R., Grothe B. (2000). Duration tuning in the mouse auditory midbrain. J. Neurophysiol. 84, 1790–1799. - PubMed
    1. Cant N. B. (2013). Patterns of convergence in the central nucleus of the inferior colliculus of the Mongolian gerbil: organization of inputs from the superior olivary complex in the low frequency representation. Front. Neural Circuits 7:29. 10.3389/fncir.2013.00029 - DOI - PMC - PubMed
    1. Cant N. B., Benson C. G. (2005). An atlas of the inferior colliculus of the gerbil in three dimensions. Hear. Res. 206, 12–27. 10.1016/j.heares.2005.02.014 - DOI - PubMed

Publication types

LinkOut - more resources

Feedback