Transferring an inborn auditory perceptual predisposition with interspecies brain transplants

Abstract

Inborn species' perceptual preferences are thought to serve as important guides for neonatal learning in most species of higher vertebrates. Although much work has been carried out on experiential contributions to the expression of such preferences, their neural and developmental correlates remain largely unexplored. Here we use embryonic neural transplants between two bird species, the Japanese quail and the domestic chicken, to demonstrate that an inborn auditory perceptual predisposition is transferable between species. The transfer of the perceptual preference was dissociated from changes to the vocalizations of the resulting animals (called chimeras), suggesting that experiential differences in auditory self-stimulation cannot explain the perceptual change. A preliminary localization of the effective brain region for the behavioral transfer by using a naturally occurring species-cell marker revealed that it is not contained within the major avian auditory pathways. To our knowledge, this is the first demonstration that abstract aspects of auditory perception can be transferred between species with transplants of the central nervous system.

Imprinting (1–4) is a developmental process through which young birds acquire knowledge about, and preferences for, physical features of their parents and species members. It is thought to operate by means of inborn perceptual preferences for stimuli sharing species characteristics, linked with specialized learning mechanisms (5–9). Although neural correlates of auditory and visual learning during imprinting have been studied well (9–16), there is little information about neural mechanisms underlying inborn perceptual preferences or their development.

Previous work using embryonic transplants of presumptive brain tissue between two bird species, the Japanese quail (Coturnix coturnix japonica) and the domestic chicken (Gallus gallus domesticus), demonstrated that it is possible to transfer inborn species differences in complex motor acts such as the production of the species-typical crowing vocalization (17–19). A naturally occurring species cell-marker (20) allows transplant localization in the brains of the resulting animals, called chimeras. Different, spatially localized regions of the presumptive brain seem to be independently responsible for the transfer of different individual components of such complex behaviors (19), suggesting independent evolutionary changes to multiple brain areas. Previous work also identified a congenital chicken–quail difference in auditory perceptual predispositions (21). This work examined responses to species “maternal calls” (Fig. 1), a well characterized vocalization given by the parents of many gallinaceous (pheasant-like) birds used to call the young to them when the chicks became too widely scattered or in the case of danger (22). Young chicks and quail preferentially responded to the call of their own species when given equal experience with the calls of both species (21).

Figure 1.

Auditory stimuli used in the approach-preference experiments. Chicken maternal call (A) and quail maternal call (B). Each call is illustrated with a sonogram (Top) showing the distribution of energy at different frequencies as a function of time (darker = more energy), and an oscillogram (Bottom) showing the amplitude variation in the calls as a function of time. kHz, kilohertz.

We were interested in determining whether, as in the case of motor behaviors, it would be possible to transplant species differences in perceptual predispositions, and if so, where in the brain the behaviorally effective regions for such transplants might reside. A total of 50 subjects (18 unoperated chickens, 15 unoperated quail, 1 chicken-donor/chicken-host chimera, and 16 quail-donor/chicken-host chimeras) were tested for individual auditory approach preferences during their first 6 days posthatching. We used the differential approach behavioral assay developed in our previous work (21) to examine subject preferences for species' maternal calls. Subjects had been hatched in incubators in the laboratory in complete isolation from all parental vocalizations; neither species emits calls during their first week posthatching that match the structure of maternal calls (refs. 23–25; E.B., unpublished observations).

During the approach tests, subjects were placed into a heated sound-attenuated chamber containing a 15 × 10 × 15 cm (length × width × height) wire cage with one recessed soft-plastic plate on each of two opposite walls. Behind the plate on each wall were a microswitch and a speaker. A digital computer connected to the box played the chicken call from the speaker on one side and the quail call from the other, at the rate of one call per 15–30 sec, and recorded the amount of time the birds spent actively pushed up against the soft plastic plate trying to get to the speaker. The computer would periodically switch the side from which each call originated. In each test session, birds would be exposed to the same number of maternal calls from both species; each subject was tested multiply during the first 6 days posthatching (21).

Because previous work in ducklings (26, 27) suggested a role for a young bird's own voice in the development of its inborn conspecific call preferences, we also monitored the sounds our transplanted animals made to ascertain whether changes in the animals' vocalizations were associated with changes in their auditory perceptual preferences.

Methods

All experimental procedures followed protocols approved by the Institutional Animal Care and Use Committee.

Chimera Production and Histological Examination.

Chimeras were produced by physically transplanting presumptive neural tissue between quail and chick embryos on the second day of incubation, as described (17). Subsequent to behavioral testing as described below, chimera brains were processed, sectioned, and stained for the visualization of the quail–chicken nuclear marker, as described (17).

Stimuli, Behavioral Testing, and Data Analysis.

Stimuli, experimental procedures, and data analyses were performed as described (21), with one added change. In the present experiments, some of the chimeric and normal subjects were implanted with testosterone on the day of hatching to induce the species-specific crowing vocalization, as described in a previous study (19). We empirically determined that the maternal call preferences of testosterone-implanted subjects do not differ from those of unimplanted subjects on either measure (individual scores: Mann–Whitney U tests, stimulus and stimulus × duration both P > 0.20, n = 18 unimplanted and 10 implanted subjects; populations of individual sessions: Mann–Whitney U test, stimulus and stimulus × duration both P > 0.25, 107 unimplanted and 64 implanted sessions). Stimuli were played and response data were collected by using the signal behavioral data acquisition and control system (Engineering Design, Belmont, MA). Recording and analysis of vocalizations were performed as described (19).

The preference index was calculated by the following formula:

1

PI is the preference index, OBS is the number of responses given to the quail stimulus, RESP is the observed number of responses in the session, PRESR is the number of times the quail stimulus was presented from the right speaker, PRESL is the number of times the quail stimulus was presented from the left speaker, PROPR is the proportion of responses given to the right speaker, PROPL is the proportion of responses given to the left speaker, N is total number of presentations, and TOT is the total number of responses given to either stimulus. Note that if a subject responds exclusively to stimuli played through one speaker, values of the PI can slightly exceed 50 if the preferred stimulus is played from that speaker fewer times than the nonpreferred stimulus.

Simulations described in the text were carried out by programs written in matlab programming language (The MathWorks, Natick, MA). We first verified that the sampling procedure produced populations of “simulated” chicken and quail individuals whose distribution did not differ from the observed distributions.

Results

Behavioral Analyses.

We analyzed the data produced by the automated approach tests for an animal's initial choice and for its choice persistence. The initial choice analysis examined which side a bird first made contact with after stimulus presentations (stimulus); the persistence analysis included all panel contact and duration of contact information (stimulus × duration). These data were corrected as described (21) to remove any biases subjects may have had toward responding differentially on the two sides of the wire cage, independently from which stimulus was being played. A mean PI was calculated (see Methods and ref. 21) to describe the degree of differential approach responses each individual subject expressed as a percentage departure from a random expectation. Negative values reflect a preference for the chicken stimulus, positive values reflect a preference for the quail stimulus, and values close to zero reflect no preference.

Fig. 2 shows the results of these tests. The degree of agreement between the stimulus and stimulus × duration measures was high (Pearson r = 0.874, P < 0.0001; Spearman ρ = 0.859, P < 0.0001; Kendall τ = 0.833, P < 0.0001; and all N = 50). Despite the fact that auditory preference behavior shows a substantial degree of variability among individuals within a species, there are highly significant population differences between unoperated individuals of the two species, as found in a previous study (21) [(means ± SE) stimulus × duration: chicken PI (Mann–Whitney U test) = −13.9 ± 4.7, quail PI = 8.9 ± 4.1, z = 2.97, and P = 0.003; stimulus: chicken PI (Mann–Whitney U test) = −12.9 ± 4.0, quail PI = 12.1 ± 3.0, z = 4.05, and P < 0.0001; chickens N = 18 and quail N = 15]. To facilitate chimera comparisons with these normative data, we adopted the following procedure. For the stimulus and stimulus × duration measures, we found the value of PI below the lowest mean quail score for an individual (Fig. 2, dashed lines), and the value of PI above the highest mean chicken individual score (Fig. 2, dotted lines). These delineations created a chicken-exclusive region (below the dashed line), a common region (between the dashed and dotted lines), and a quail-exclusive region (above the dotted line).

Figure 2.

Simultaneous auditory choice test results. PI values are shown for stimulus (Left) and stimulus × duration (Right) measures of all subjects. (○), mean values for individual unoperated chickens; (●) mean values for unoperated quail; (□) chimeric animals. Fill color corresponds to type of surgery performed, illustrated on schematic embryonic neural tubes (Center). (Bars = ±1 standard deviation.) CC, chicken-donor, chicken-host rostral mesencephalon/caudolateral diencephalon (M1 type) surgery; D, quail-donor, chicken-host rostral diencephalon surgery; F, quail-donor, chicken-host forebrain/rostromedial diencephalon surgery; H, quail-donor, chicken-host rostral hindbrain surgery; HM, quail-donor, chicken-host rostral hindbrain/caudal midbrain surgery; MF, quail-donor, chicken-host rostral midbrain/diencephalon/forebrain surgery; M1 and M2, quail-donor, chicken-host rostral mesencephalon/caudolateral diencephalon (M2 smaller than M1).

Identification of Chimeras with Altered Auditory Perceptual Preferences.

The 16 quail–chicken chimeras represented seven types of surgeries involving different parts of the neural tube (shown schematically in Fig. 2). The only surgeries that produced individuals with preference scores in the quail-exclusive region on one or both measures were the three types of operations including the presumptive rostral midbrain and caudolateral diencephalon (MF, M1, M2, Fig. 2). If we consider the seven individuals with these surgeries as a group, their distribution between the quail-exclusive region and the region where chickens are found is significantly different from the normal chicken distribution for both response measures [stimulus × duration: chicks, 18 common region, 0 quail region; rostral midbrain chimeras: 3 common region, 4 quail region, P = 0.0028 (Fisher's exact test); stimulus: chicks, 18 common region, 0 quail region; rostral midbrain chimeras: 1 common region, 6 quail region, P = 0.0001 (Fisher's exact test)], and not significantly different from the distribution of normal quail on either measure [stimulus × duration: quail, 12 common region, 3 quail region; rostral midbrain chimeras: 3 common region, 4 quail region, P = 0.1447 (Fisher's exact test); stimulus: quail, 6 common region, 9 quail region; rostral midbrain chimeras: 1 common region, 6 quail region, P = 0.3501 (Fisher's exact test)]. These results demonstrate a related group of surgeries for which there has been a transfer of an inborn auditory predisposition by means of embryonic neural transplants.

These seven animals had a mean stimulus × duration PI value of 19.03 ± 4.64, and a mean stimulus PI value of 20.23 ± 2.61, considerably higher than the quail mean values of 8.9 ± 4.1 and 12.1 ± 3.0, respectively. A simulation test was run in which 1,000 “simulated” quail were created by randomly picking seven testing sessions (the mean number per subject) from the observed distribution of unoperated quail sessions and taking their mean. We then drew 1,000 “groups” of seven birds each from the distribution of simulated quail individuals, calculated their mean PIs, and recorded the number of times out of 1,000 in which this mean was greater than or equal to the observed mean for the group of seven chimeras, giving a probability that these results could be observed by chance. This procedure yielded a value of 0.017 for the stimulus × duration result and 0.006 for the stimulus result, suggesting that this group of chimeras had a significantly more extreme quail preference score than the population of unoperated quail.

A control transplant involving the same part of the rostral mesencephalon/caudolateral diencephalon between two individual chickens (CC in Fig. 2) had a preference score on both measures well within the normal chicken range, as did quail–chick chimeras with transplants in the rostral diencephalon (D), in the hindbrain (H), and in the rostral hindbrain/caudal midbrain (HM). The four animals in the forebrain/rostromedial diencephalon (F) transplant group had a distribution of measures on the stimulus score that looked no different from normal chickens, yet their scores on the more sensitive stimulus × duration measure looked intermediate between chickens and quail. The distribution of these scores between the chicken-exclusive region and the common/quail region is not different from that of unoperated chickens [chicken: 8 chicken region, 10 common/quail region; F chimeras: 1 chicken region, 4 common/quail region, P = 0.6106 (Fisher's exact test)]. But when we applied a simulation test similar to the one described above (generating a “simulated” population of 1,000 chickens, and then randomly sampling groups of four birds from this population 1,000 times to see the probability of generating 3 of 4 subjects with preference scores greater than zero), the procedure yielded a value of 0.024 for the stimulus × duration result. This value suggests that there may be a very subtle shift in the behavior of some of these subjects— although they have a strong tendency to respond to chicken sounds first, when they do respond to quail sounds they tend to do so for longer durations than unoperated chickens. This outcome is potentially caused by chicken and quail cells mixing near the transplant boundaries in the regions of the brain that contribute to the shift in auditory preference (see below).

Auditory Perceptual Transfer Is Independent of Species' Vocal Behavior.

We monitored the sounds that the chimeras with altered auditory perceptual preferences made (MF, M1, and M2 transplants). We were able to witness and/or to record four of these seven subjects giving peeps, isolation calls, and testosterone-induced juvenile crowing vocalizations (crowing was recorded in three of these cases). Although the peeps and isolation calls were not noticeably different from those of normal chickens, these vocalizations are variable during the first week of life (25), and, once differences in body size are taken into account, may be difficult to reliably discriminate in chickens and quail (E.B., unpublished observations). Crowing is a more rigorous assay, because of well documented species differences and previous work showing that it can be transplanted from quail into chickens (17, 19). In all four cases recorded or witnessed, crowing was indistinguishable from that found in normal chickens in MF, M1, and M2 animals (Fig. 3). This result demonstrates a dissociation between the auditory preferences and the vocal behaviors of the chimeras, and suggests that differences in auditory self-stimulation are unlikely to be responsible for the change in auditory preferences seen in MF, M1, and M2 transplants.

Figure 3.

Sound spectrograms of testosterone-induced juvenile crowing vocalizations. Two unoperated quail (Top), two chimeras with preferences in the quail-exclusive region on both behavioral measures (Middle), and two unoperated chickens (Bottom). (Left) FM chimera. (Right) M2 chimera. Unoperated chicken examples were chosen from a large collection based on matching temporal properties with the chimeras' crows. FM and M2 as in Fig. 2. kHz, kilohertz.

Preliminary Localization of the Effective Region for Auditory Perceptual Transfer.

We used the quail–chicken nucleolar staining difference (20) to localize preliminarily the brain region that seems to be necessary and sufficient for evoking the quail auditory preference in chickens. Here, the brains of three M1/M2 chimeras, whose behaviors were in the quail-exclusive region for both measures, and one F chimera, whose behavior was in the chicken-exclusive region for both measures, were especially useful. Fig. 4 (dark shading) indicates the consensus region shared by animals with behavioral scores in the quail-exclusive region and not shared by animals in the chicken-exclusive region or with behavioral scores below zero. This region runs bilaterally between anterior 4.2 and 6.6 in the chick brain atlas of Kuenzel and Masson (28) (coordinates in Table 1). It is especially notable that major areas of the avian primary auditory system such as the MLd/ICo/Torus complex (29) (thought to be homologous to the inferior colliculus of mammals), and the nucleus ovoidalis (28) (the major thalamic auditory relay projecting to the telencephalon), lie outside of this region. Close examination of the three F chimeras with stimulus × duration scores greater than zero revealed that two of them had scattered quail cells that fell within this region bilaterally.

Figure 4.

Preliminary localization of behaviorally effective transplant region (dark shading), indicated schematically on a three-dimensional view of a chicken brain. Background arrow indicates a line going from anterior (A) to posterior (P). Telencephalon (left) and cerebellum (center) are represented as transparent areas (dashed outlines). Dashed lines also represent the position of field L (FL), the primary telencephalic auditory recipient area. The darkly shaded region includes nucleus pretectalis (PT, shaded) and part of nucleus rotundus (ROT, partially shaded) but not the thalamic nuclei dorsolateralis anterior thalami (DLA) or ovoidalis (OV) (both unshaded). The area ends anterior and medial to the MLd/Ico/Torus complex (MLd, unshaded). nVIII, Cochlear nucleus.

Table 1.

Approximate coordinates for the behaviorally effective transplant region from the chick brain atlas of Kuenzel and Masson (28)

Anterior	Mediolateral	Dorsoventral
4.2	1.0–2.5	0.5–6.0
4.4	2.0–3.5	0.5–6.0
4.6	1.5–4.0	0.5–5.5
4.8	2.5–4.5	0.5–5.5
5.0	1.5–4.0	0.5–5.0
5.2	2.0–4.0	0.5–4.5
5.4	2.0–4.5	0.5–4.0
5.6	2.5–4.5	0.5–4.0
5.8	2.5–4.5	0.5–3.5
6.0	2.5–5.0	0.5–3.5
6.2	2.5–6.0	0.5–3.0
6.4	2.5–4.0	0.5–2.5
6.6	2.0–4.0	0.5–2.5

Units are in mm. 

Discussion

The results of this study collectively demonstrate a neurobiological basis for species differences in an inborn auditory perceptual predisposition, independent of an animal's auditory or vocal experience, and suggest a region of the brain that is crucially important for the development of these predispositions. As in previous work with inborn differences in complex motor acts (19), these results could be explained either by changes to an autonomously operating bilateral “brain module,” which detects and signals the presence of an “interesting” stimulus to the rest of the brain, or by developmental effects of the transplanted region on host regions of the brain, whereby transplanted (genetically quail) cells affect emerging phenotypic circuit properties of genetically chicken cells that they form connections with.

It is especially intriguing that the population of chimeras with transformed auditory preferences had more extreme behavioral scores than normal quail, suggesting some kind of amplification of behavioral effect. This phenomenon implies a functional or developmental interaction through which quail cells disproportionately influence circuit behavior. Two possible mechanisms include (i) differences in developmental timing between the chicken and quail parts of the circuit(s). If the quail cells provide developmental and/or functional influences earlier (and therefore over a longer time interval in the host embryo) than their chicken counterparts, quail cells could have a competitive advantage in driving the behavior of the cells with which they connect. (ii) Quail cells may make more efficacious functional synapses because of some difference in the density of their projections, the structures of their synapses or arborizations, or the timing patterns of their action potentials. Each of these factors singly or in combination may result in quail cells functionally dominating the behavior of the circuit(s) they are in.

Such phenomena may become more clearly understood in the future by a precise localization of the effective transplanted region, and by studying its connections with the auditory system and its developmental interactions with other brain regions it forms connections with. Examining the development of the phenotypes of host cells that are connected with donor cells in the transplants will be especially important to this endeavor. We are currently unaware of any studies demonstrating auditory activity within the transplanted region, or connections linking this region to known auditory areas. Elucidating the functional role of the effective transplanted region in the auditory system, and its counterparts in other vertebrate species, may provide new insights into the dynamics of auditory perceptual development.

This work also demonstrates the significant role played by central brain pathways in the determination of global auditory perceptual properties relevant to attention and learning during early postnatal development, as well as the role that changes to these same pathways play in the evolution of species perceptual differences. Both the sites of action and the developmental mechanisms used to achieve such evolutionary changes in higher auditory function are now amenable to direct empirical investigation.

Abbreviation

PI

preference index