NEUROETHOLOGY OF CRYING
, Head, Unit on Developmental Neuroethology
Michelle Becker, PhD, Guest Researcher
Deborah Bernhards, BS, Biological Technician
Catherine Depeine, BS, Technical Training Fellow
William Kenkel, BA, Technical Training Fellow
Karen Ladd, BS, Summer Student
Crying is a universal mammalian behavior in infancy and an essential signal that activates care-giving behavior. Little is known about the neural basis of crying or about why crying can be such a compelling stimulus to the listener. Our goals are to (1) determine the neural pathways that underlie cry production and cry perception; (2) track the developmental course of crying in infancy, particularly to determine the roles of inheritance and experience in individual variability in crying behavior; and (3) examine the interplay of cry acoustics and the hormonal and experiential status of care-givers in regulating the motivation to respond to a crying infant. Our approaches include behavioral experiments in nonhuman primates aimed at defining the critical features of infant crying that promote care-giving; acoustic analysis of cry sounds in the search for acoustic markers of developmental status, familial traits, environmental influences, and neurological risk factors; and functional neuroanatomical studies aimed at defining the neural populations activated during crying and cry perception.
The cry circuit in the brain
A necessary step in understanding how the brain regulates crying and cry responding is to identify the neural circuits that mediate the two behaviors. We studied the brains of infant common marmosets by using immunocytochemical identification of Fos, the protein product of the immediate-early gene c-fos, as a marker for functional activity of neurons following an extended bout of crying. We had demonstrated earlier that Fos is a reliable marker of functional activity of neurons in adult monkeys; now, for the first time, we are using Fos to studying the brains of infant monkeys. Using a Leica microscope with a motorized stage and the Life Science program from Bioquant, we are systematically digitizing thousands of images at 100x magnification from sections throughout the brains of monkeys at 1, 2, 3, and 4 months of age. We subsequently collect the images into montages of each section, quantify the number of Fos-expressing neurons per image, and create a map of the distribution of regions within each montage with the greatest number of Fos-expressing neurons. We stain adjacent sections for Nissl granules, allowing us to make detailed anatomical identification of the regions of greatest expression. Reference to a brain atlas produced in our laboratory assists in the construction of a “wiring diagram” of the structures making up the circuit underlying cry production at different ages during development.
New findings this year derive from the digitization and measurement of Fos-labeled cells from six infant brains, two at 1 month of age, two at 2 months of age and two at 4 months of age (including one control brain). We digitized approximately 330 to 620 images (depending on the size of the original section) from each section and counted the labeled cells in each image. We created a grid overlay for each montage and entered the number of labeled cells from each image into the appropriate grid space; the result was an XY map of the distribution of counts for each montaged section. Examination of the map for each section clearly showed a differential distribution of subregions with high numbers of labeled cells, including periventricular grey in the diencephalon, septum, preoptic area, amygdala, and hippocampus in the rostral forebrain, anterior cingulate gyrus and gyrus rectus in the frontal cortical regions, superior temporal gyrus, and periaqueductal grey of the midbrain. From studies of adult nonhuman primates, most of these regions are known to be involved in auditory communication, but our study is the first to demonstrate that the vocal production system is widely activated in such young monkeys. High counts in the superior temporal gyrus presumably reflect neurons in the auditory cortex activated during vocalization by the infant. We are currently counting additional sections from this group of infants and are developing a false-color map that will convert the numbered grids into areas of varying color indicative of the relative intensity of activity in each subregion of the infant brain.
This year, we began to identify neurons in the brains of marmosets listening to infant cries. For the initial study, we exposed five adult marmosets to 30 minutes of recorded infant cries; we then euthanized them and processed their brains for Fos immunocytochemistry. A preliminary analysis of three brains (including one control animal not presented with infant cries) demonstrated, for the first time in a nonhuman primate, that, in addition to expected activation of temporal lobe auditory areas, infant cries activate several defined areas of the brain, including the anterior cingulate gyrus and midline frontal cortex rostral to the expected area and hippocampus.
Marmoset brain atlas
To assist in describing the neuroanatomy of the cry circuit, a brain atlas can guide identification of the structures under examination. Two brain atlases are available for the squirrel monkey and another for the New World primate, but the only atlas published for the marmoset has been out of print for many years. Therefore, we undertook the task of creating a marmoset brain atlas. With a digital camera, we photographed frontal (coronal) sections cut through the brain of an adult common marmoset and then stained them for cell bodies (Nissl stain) and fiber tracts (Weil stain); we used the resultant images for the atlas. In particular, Rachel Bell used one of the existing squirrel monkey atlases as an aid in identifying and labeling the corresponding structures on the marmoset brain sections. To improve the usefulness of our atlas as a reference, we created schematic images of the photographed sections with Adobe Photoshop and the NIH Image-J program. We labeled the schematic images rather than the photographs and used a tablet monitor with an electronic stylus to create wire-frame tracings of the brain sections. The tracings lend themselves to stretching or other manipulation for overlaying MRI images in planned studies of marmosets that will use the MRI and functional MRI techniques. A published atlas is planned and will contain both the photographs and the labeled schematic images. In addition, we stained for Nissl granules and then imaged an extensive series of frontal brain sections from infant marmosets.
We employed the brain atlas (created from the brain of an adult female) for the first time to identify structures in the brains of infant marmosets. We prepared six infant brains for histological examination by staining sections with cresyl violet in order to reveal the distribution of Nissl granules. We examined two brains each from 1-, 2-, and 3-month-old infants. We found that, even in the brains of 1-month-old infants, we could identify at least 20 structures found in the adult. The major difference between infant and adult brains was the absence of large myelinated tracts in the former, making it difficult to discern the boundaries of some structures, particularly in the thalamus.
Cry characteristics of infant marmosets
We used Raven, a software program developed at Cornell University, to measure the acoustic structure of the cry sounds of infant marmosets. We recorded cries when infants were separated from their family group for short periods, which is when extensive crying typically occurred. We then digitized the recordings and stored them on a computer for analysis. We generated a sound spectrogram of each cry on a computer screen and used a computer mouse to save the time and frequency values for four pre-selected points on each sound spectrogram. Given that cries typically occur in bouts of two, a pair of cries produced eight time and frequency values. From these values, we used macros created by Karen Ladd in Microsoft Excel to calculate a total of 34 acoustic parameter measures (in the frequency and time domains) for statistical analysis. We recorded a total of 3,162 cries from infants at 2 to 3 months of age, from offspring that had reached maturity (18 months to 2 years of age) and from the parents of the infants, and then digitized and subjected the cries to statistical analysis. Data from two unrelated groups (colonies) contributed to the study. We are analyzing the degree of individuality (vocal signature) in marmoset cries in adulthood and during infancy; whether the cries of littermates (fraternal twins) resemble each other more than the cries of age-matched unrelated individuals; whether familial traits transcend specific litters (i.e., shared acoustic characteristics of offspring born to the same set of parents); and the extent to which the cries of infants resemble the cries of their parents (i.e., calls made by the parents when they are briefly separated from their family groups). In addition, we wish to determine the developmental age at which sex differences in cries occur (a phenomenon in adults discovered by our laboratory) as well as the age at which cry syntax (the relationship between the structure of the first and second element of a two-cry series) emerges. We found that most individuals, infants as well as adults, demonstrated strong individuality (as shown by discriminant function analyses). We also found litter differences for nearly all litters and noted that sex differences in cry structure emerged as early as two months of age and that syntax in cry pair structure may be demonstrated at two months of age. These findings indicate tight regulation of marmoset cry structure starting early in infancy, suggestive of strong genetic control.
This year, we employed, for the first time, another program called Syrinx to annotate digitized sound stream files. Catherine Depeine, who had experience with Syrinx, annotated 20-minute sound streams (collected from vocalizing infants and older individuals separated for 20 minutes from their family group) and tested the consistency of vocal behavior over this length of time. She found no significant difference in the rate and type of calls between the first 5 minutes and last 5 minutes of the 20-minute period. The results suggest that the 5-minute period that we typically use for infants is adequate to characterize crying at a given age, thus minimizing stress on the separated infant.
A paradigm for measuring cry responsiveness
One of the great mysteries of parenting is why the cries of an infant are so compelling in activating care-giving behavior. In humans, it has long been assumed that the hormonal state of the mother plays a role in activating and maintaining infant retrieval and nurturance. However, fathers also play an important role in infant care. The same holds true in the common marmoset. In addition, older brothers and sisters exhibit a great deal of interest in infants and gain care-giving experience by assisting in carrying the infants when they are not being fed. We wish to understand the mechanisms of infant care, particularly with respect to the role that infant crying plays in care-giving. To this end, we developed a testing paradigm for evaluating subject interest in infants and infant crying. The paradigm is based on the orienting and approach behavior exhibited by a subject in a Y-maze when presented with a stimulus at the end of one of the arms. We began using this testing paradigm last year and have added more test trials this year. We employ two stimulus paradigms: a live infant and an audio speaker emitting recorded cries from an infant. In the tests using live infants, we brought subjects into the test room in a small transfer cage and placed them at the base of the Y-maze after we had placed a live infant at the end of one of the arms. After 15 seconds, we raised the door to the transfer cage and evaluated the subject’s behavior. We defined a positive response as a parent spending more time in the arm containing the infant than in the other (empty) arm. Of eight adults with parenting experience, seven gave positive responses in at least one of two trials. To date, we have run 72 trials with a live infant and 50 trials with cries presented via speaker. We also tested some individuals with no stimulus (control) to assess general activity in the Y-maze. Out of 72 trials with a live infant, 27 yielded a positive response (pass). All eight parents received a pass score on at least one trial, with one parent scoring a pass on 4 of 5 trials. We also tested juvenile offspring of these parents, but they performed considerably more poorly. Several juveniles failed to achieve even one pass score. However, the female offspring of the parent with the most pass scores also received a pass in three out of three trials. Tests involving recorded cries presented via speaker yielded 17 pass scores. All parents received at least one pass score. The parent with the most pass scores in a test with a live infant also produced the most pass scores with recorded cries. Although tested on only one trial, the offspring with the most pass scores with a live infant also achieved a pass score with the recorded cries. We plan to move the testing apparatus to our main primate holding building and then test the rest of the colony. In addition, in an attempt to correlate home-cage behavior with behavior in the formal testing procedure, we hope to tabulate the amount of carrying-time of subjects in their home cages with their own offspring or siblings.
Comparative studies of crying infants
Catherine Depeine digitized an extensive database of the cry sounds of the infants of several prosimian primates, collected at the Duke University Lemur Center, and organized them into acoustic subtypes. We were interested in evidence that the infants of nocturnal prosimians would produce a different subtype of separation-induced vocalization. Acoustic analysis of the vocalizations showed that the infants of nocturnal primates do produce a markedly different type of call, best characterized as a “click” sound. We hypothesized that the click sounds are adaptive in that they provide additional localization cues (clicks are more readily localizable than tonal sounds) and are useful in the dark for mothers locating their separated infants. The mechanisms that mediate the click-like sounds, at both at the laryngeal and brain levels, are unknown but represent a possible area of future study.
Newman JD. Neural circuits underlying crying and cry responding in mammals. Behav Brain Res 2007;182:155-65.
Soltis J, Wegner FH, Newman JD. Urinary prolactin is correlated with mothering and allo-mothering in squirrel monkeys. Physiol Behav 2005;84:295-301.
1 Rachel Bell, former Summer Student
COLLABORATOR
Nicholas Bock, PhD, Laboratory of Functional and Molecular Imaging, NINDS, Bethesda, MD
Afonso Silva, PhD, Laboratory of Functional and Molecular Imaging, NINDS, Bethesda, MD
For further information, contact newmanj@mail.nih.gov.
