Coding interaural time differences at low best frequencies in the barn owlIntroductionMaterials and methodsSurgeryStimulus generation and calibrationStimulus paradigms and unit characterizationResultsBehavioral dataPhase-locking below 2 kHzAnatomy and cell types of the low frequency pathway in owls: delay lines at low best frequencies?Recordings from low CF nucleus laminaris neurons: coincidence detection below 2 kHzDiscussion and conclusionsReview of temporal coding in birds and mammalsPresynaptic specializations for encoding temporal informationPostsynaptic specializations for encoding temporal informationTemporal precision in responses of the cochlear nuclei and nucleus laminarisSensitivity to interaural time differencesDelay line-coincidence detection circuits in birdsCoincidence detectors in birds and mammalsStereausis and delay linesEncoding ITDs at low best frequencies in birds and mammalsAcknowledgementsReferencesCoding interaural time differences at low best frequenciesin the barn owlCatherine E. Carra,*, Christine K€opplbaDepartment of Biology, University of Maryland, Biology–Psychology Building, Room 4227, College Park, MD 20742-4415, USAbLehrstuhl f€ur Zoologie, Technische Universit€atM€unchen, Lichtenbergstrasse 4, 85747 Garching, GermanyAbstractIn birds and mammals, precisely timed spikes encode the timing of acoustic stimuli, and interaural acoustic disparities propagateto binaural processing centers. The Jeffress model proposes that these projections act as delay lines to innervate an array ofcoincidence detectors, every element of which has a different relative delay between its ipsilateral and contralateral excitatory inputs.Thus, interaural time difference (ITD) is encoded into the position of the coincidence detector whose delay lines best cancel out theacoustic ITD. Neurons of the avian nucleus laminaris and mammalian MSO phase-lock to both monaural and binaural stimuli butrespond maximally when phase-locked spikes from each side arrive simultaneously, i.e. when the difference in the conduction delayscompensates for the ITD. McAlpine et al. [Nat. Neurosci. 4 (2001) 396] identified an apparent difference between avian andmammalian ITD coding. In the barn owl, the maximum firing rate appears to encode ITD. This may not be the case for the guineapig, where the steepest region of the function relating discharge rate to interaural time delay (ITD) is close to midline for all neurons,irrespective of best frequency (BF). These data suggest that low BF ITD sensitivity in the guinea pig is mediated by detection of achange in slope of the ITD function, and not by maximum rate. We review coding of low best frequency ITDs in barn owls andmammals and discuss whether there may be differences in the code used to signal ITD in mammals and birds.Ó 2004 Elsevier Ltd. All rights reserved.Keywords: Auditory; Bird; Brainstem; Evolution; Timing1. IntroductionIn the auditory system, accurate coding of temporalinformation has direct behavioral relevance for soundlocalization. In birds and mammals, precisely timedspikes encode the timing of acoustic stimuli, and inter -aural time differences (ITDs) propagate to binauralprocessing centers such as the avian nucleus laminarisand the mammalian medial superior olive [20,47,130].In birds, projections from the cochlear nucleus mag-nocellularis to the nucleus laminaris act as a delay lines[19,82] and the projection from mammalian sphericalbushy cells to the medial superior olive has been pro-posed to serve a similar role [6,107]. Delay line axonssynapse on an array of coincidence detectors, everyelement of which has a different relative delay betweenits ipsilateral and contralateral excitatory inputs. Thus,ITD is encoded into the position (a place code) of thecoincidence detector whose delay lines best cancel outthe acoustic ITD (for reviews, see [47,53]). The neuronsof nucleus laminaris and MSO act as coincidencedetectors, in that they phase-lock to both monaural andbinaural stimuli and respond maximally when phase-locked spikes from each side arrive simultaneously, i.e.when the difference in the conduction delays compen-sates for the ITD [20,36,82,86,87,128]. In this theory, anetwork composed of delay lines and coincidencedetectors explains how ITDs are computed [44].The coincidence detectors of nucleus laminaris andMSO fire maximall y at a particular ITD (and at inter-vals 2p apart). Recordings from these neurons showmultiple peaks at time lags corresponding to the ITDand integer multiples of the stimulus period (see [54]). Inspecific examples from dog [36], cat [128] and barn owl[20], this preferred ITD ha s been shown to be equal andopposite to the difference in neural delays between theresponses to stimulation of either ear alone. Measures of*Corresponding author. Tel.: +1-301-405-2085; fax: +1-301-314-9358.E-mail address: [email protected] (C.E. Carr).0928-4257/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.doi:10.1016/j.jphysparis.2004.03.003Journal of Physiology - Paris 98 (2004) 99–112www.elsevier.com/locate/jphysparisthis characteristic delay in cats and rabbits [5,127] andbarn owls [119] show a coincidence of peaks in theinferior colliculus, and in the nucleus laminaris [20,87].It seems likely that barn owl and cat IPD coding neu-rons signal the best ITD or characteristic delay througha peak or rate code mechanism.Recent studies in the guinea pig [71] have pointed outthat there are major difficulties associated with using apeak code to signal ITD at low best frequencies. Thesedifficulties are particularly acute when the anima l has asmall head and only a few hundred microseconds ofavailable interaural time difference. McAlpine et al. [71]have shown that at frequencies below abou t 1 kHz, it isthe slope of the ITD plot that tends to be centeredaround 0° ITD, not the peak. Is there a fundamentaldifference between birds and mammals in the code usedto signal ITD in low frequency sounds, or do both birdsand mammals use the slope to determine location at lowbest frequencies, and switch to peak detection when theinformation content of the peak increases?We will review the literature on ITD coding in birdsand mammals and present a few recordings from lowbest frequency barn owl NL that suggest that low bestfrequency coincidence detector neurons in the owl alsohave the slope of the ITD function within the biologicalrange of ITDs, and not the peak. Both sections will beused to support the