Response plane histograms are shown for 2 units (A,C,E and B,D,F) with different best periodicities (600 and 900 Hz, respectively) and hence located in different areas of the circular periodicity map in AI (as schematically indicated by arrows from the periodicity map between left and middle column; numbers in map refer to BPs of the respective map area) which were recorded simultaneously during stimulation with either a single AM tone complex (A,B) or 2 concurrently presented AM tone complexes (C–F). Response plane histograms are arranged in panels where modulation frequency of AM tone complexes (in steps of 150 Hz) is plotted over time. Spikes per bin are color-coded according to the scale bar at the bottom right of each histogram, with the maximum number of spikes indicated in each case. Histogram binwidth: 5 ms. Vertical yellow lines mark onset and offset of stimulation. In the 2 AM tone condition in C,D, the best AM tone complex for each unit was fixed, while the concurrent AM tone complex was varied in fm (cf. Fig. 1F). In E,F the fixed AM tone complex was that of the other unit, that is, C,F and D,E (and A,B) were recorded simultaneously. The unit in A,C,E (the ‘loser’) responded only to the two AM tone complexes when the periodicities of both AM tone complexes fell within its periodicity receptive field. By contrast, the unit in B,D,F (the ‘winner’) always responded to its best AM tone complex irrespective of the periodicity of the concurrent AM tone complex. Note that the strength of the response to the best AM tone complex could be modulated by the second AM tone complex: Strongest responses were typically seen, when both AM tone complexes had similar periodicities, that is, when both periodicities were within the receptive field of the unit.

Responses plane histograms in response to two concurrently presented AM tone complexes are shown for the same unit in AI before (A), during (B), and after (C) iontophoretic application of BIC; ejecting current: 40nA. Conventions and layout as in Fig. 2C,D. The unit showed a ‘loser’-like response behavior in the pre-drug condition, but a ‘winner’-like response behavior during blockade of GABA_A-mediated inhibition via iontophoresis of BIC.

Panels A and B show axonal fibers and terminal boutons labeled anterogradely following injections of biocytin into AI of the gerbil. Panels C–E demonstrate direct contacts of biocytin-labeled terminations with parvalbumin-positive (i.e. GABAergic) interneurons. A: Single sagittal section through the auditory cortex processed for biocytin following an injection of biocytin (100 nl) into layers II–IV of AI. Labeled axons (stained brownish-yellow) expand from the injection site (red star, BF 1–2 kHz) to a large extent in dorsal direction, i.e. in a tonotopic manner within the dorsoventrally oriented 1–2 kHz isofrequency contour (schematically indicated by grey solid lines). Additionally, a substantial number of axon projections cross different frequency domains (i.e. are non-tonotopic), in this case predominately in ventrorostral and ventrocaudal directions (red arrows). Note, that these non-tonotopic, long-range projections spread asymmetrically from the injection site (indicated by different length of red arrows). Additionally, projections to the anterior and posterior auditory fields (AAF and DP) can be seen. B: Reconstruction of five consecutive sagittal sections through supragranular and granular layers of the gerbil's auditory cortex following an injection of biocytin (5 nl) into layers III–IV of AI (red star; BF 4 kHz). Red lines represent outlines of traced sections, green-shaded area corresponds to the auditory koniocortex (which comprises fields AI and AAF [14], [28]). Blue stars represent biocytin-filled neurons at the injection site (shown on microphotograph in the inset). Blue dots represent biocytin-labeled boutons. Note again the asymmetric distribution of labeled boutons of the non-tonotopic, long-range projections within AI (red arrows). C–E: Biocytin-labeled terminations (brownish stain) in contact with various PV-positive GABAergic interneurons (violet stain) in layers III/IV of AI. Arrows point to synaptic contacts at proximal dendritic aspects (C, D) and at cell somata (E). The distance of contacted PV-positive somata from the injection site is noted in the top right-hand corner of each panel. Section orientation in A applies to all panels. The experimental cases from which data are shown as well as the volume of biocytin injected are noted in the bottom left-hand corner of each panel.

The concept of local cortical (lateral) inhibition (left) is contrasted with that of global cortical inhibition (right), as illustrated by examples of a tonotopic and a periodicity map obtained by optical imaging in gerbil AI (A,B, cf. Ref. [17]) and by schematic drawings of the interconnectivity pattern (C,D). A: Different colors within the tonotopic map depict representations of different pure tone frequencies from low (blue) to high (red). B: Different colors within the periodicity map depict representations of different AM tone periodicities from low (blue) to high (red). The concept of local (lateral) inhibition proposes an inhibitory interconnection pattern, whereby any given unit (or stimulus representation) inhibits only its immediate neighbors within the parameter space (A,C: red projections), resulting in local contrast enhancement. The concept of global inhibition proposes an inhibitory interconnection pattern, whereby any given unit (or stimulus representation) inhibits all other representation within the parameter space (B,D: blue projections, inhibitory interneurons are not shown), resulting in the implementation of a “winner-take-all” algorithm, i.e. global contrast enhancement. Note that since any BP representation in a circular map has an eccentric location, projections from any BP representation within the map to all other locations of the map (blue lines) result in an asymmetric geometry of projections (cf. Fig. 6B). Gray lines: inactive projections. Black arrows: Thalamic input. E,F: Stimulation paradigms (schemes) used to test local (red) and global (blue) inhibitory concepts, respectively. E: Isointensity frequency response functions (light red curve) are usually obtained by plotting pure tone (vertical dark red line) evoked discharge rates as a function of tone frequency. F: Stimulation paradigm used here in the competitive interaction experiment. Two AM tone complexes were presented simultaneously (duration: 200 ms, 65 dB SPL). Spectra of both AM tone complexes (vertical dark blue lines) were entirely outside the frequency receptive field (FRF; light blue curve; cf. Ref. [33]) of the unit, which in our experiments always meant above the FRF, because all recorded units showed responses to low frequency pure tones. One of the complexes had a fixed fm (fc1/fm1) set to best periodicity of one of the units, the second had a different fc (fc2) and varied in fm (fm2). Note that fc2 could be higher or lower than fc1 in the experiments, but both AM spectra were always completely above the units' FRF.

(A) Relationship between response range and range of presented stimuli. By definition, winner units respond to all stimulus combinations in two AM conditions. (B) Distribution of tuning sharpness (Q) during two AM conditions. Tuning of losers is significantly sharper than tuning of winners. Depending on BP, Q-values of winners may only range between 0 and 1, whereas those of losers may exceed 1. (C) Comparison of tuning sharpness (Q) between one and two AM conditions. Whereas there is no significant change in Q in losers, winners show significantly smaller Q-values during the two AM condition compared to the one AM condition. (D) Frequency distribution of tuning filter characteristics. Complex (CX) filter characteristics were most frequent in both winners and losers, but whereas only winners showed high-pass (HP), non-selective (NS) and band-suppression (BS) filter characteristics, band-pass (BP) and low-pass (LP) filter characteristics were only found in loser responses.

(A) Relationship between response range and range of presented stimuli whereas some units (blue) changed their response properties from loser to winner type, others stayed losers (red) but even then typically responded to a wider range of stimulus combinations (red dots above diagonal). Winners (pink) always stayed winners under drug conditions. (B) BIC-effect on tuning sharpness. Q-values became significantly smaller in those units that changed from loser to winner response type (blue), but this effect was only a trend in the group of losers that stayed losers (red). Winners (pink) also showed significantly smaller Q-values during the BIC-condition because of a drop of BP. (C) A similar effect was seen with a modified Q-value where the BP of the control condition was also used to calculate the Q-value of the BIC-condition. With this analysis consequently tuning sharpness of winners did not change at all (pink).