Background: An infant's cry is an innate, universal human behavior which seemingly emerged in the evolutionary transition from reptiles to mammals, simultaneous with the onset of strong maternal behavior and a well-developed limbic system (1). Mammalian lesion data suggest that the cingulate gyrus, a phylogenetically newer part of the limbic system, is involved in both infant crying and maternal behavior (See (1) for a review). Here, we report on the feasibility of using BOLD-fMRI to understand brain activity in human mothers listening to infant crying and hypothesize that this auditory stimulus will produce cingulate activation, especially anteriorly. We also comment on variables we are encountering in designing a more rigorous follow-up study.

Methods: Subjects were 4 biological, premenopausal  mothers having a youngest child less than age 4 (range 3 weeks- 3 1/2 years). BOLD-fMRI imaging was performed in a Picker Edge 1.5T MRI scanner equipped with high performance whole-body gradients. T1-weighted structural and T2*-weighted echoplanar, functional images were acquired for 15 coronal slices (TE 40, TR 3000, FOV 31.5 cm, 5mm/skip2).  During the task, subjects listened to 30 second baby cries (active condition) alternating with 30 second white noises (rest condition) made by replacing cries with white noise bursts to simulate the normal temporal pattern of crying. Both pain and separation cries were used. After imaging, subjects rated their emotional reactions during the cries and white noises. We then analyzed functional data using a paired t-test comparing all active minus rest scans (P < 0.001). The first and last authors assessed the following areas for mild, moderate, and robust activity within each subject: anterior, middle, and posterior cingulate (AC, MC, and PC), right and left insular cortex (RI, LI), orbitofrontal cortex (ROF, LOF), dorsolateral prefrontal cortex, temporal poles, medial temporal lobes (RMT, LMT), caudate (RCA, LCA), thalamus (RTH, LTH), and brainstem (BS). We labeled areas activated if they displayed at least moderate activity.

Results: All subjects felt more sad and experienced more of an urge to help during the cries than during the white noises. Other emotional reactions were variable. Two of 4 subjects had AC, 1 of 4 had MC, and 1 of 4 had PC activation. Three of 4 subjects had ROF activation. Other areas were activated in only 1 of 4 subjects including RI, LMT, RMT, LCA, LTH, and BS. (see Figure 1)

Significance:  This small pilot experiment demonstrates the feasibility of studying maternal response to infant cries in a fMRI scanner environment. It tentatively supports the notion that the cingulate is involved in response to infant crying. Consistent with some monkey lesion work, it suggests that the ROF may also be involved (2). In designing follow-up studies, we are considering such factors as maternal and infant temperament, the type of infant cry, using each mother's infant's own cries, the appropriateness of our control stimulus, postpartum timing, and allowing a mother to actively terminate the cries as in natural settings. Future work in this area may lead to understanding the brain basis of mother-infant interaction and the biological roots of child neglect and abuse.
 

Figure 1: Anterior (left) to posterior coronal images showing ROF and AC activation (P<0.001) in a subject.
 

(1) MacLean, PD.  The Triune Brain in Evolution, 1990, Plenum Press: New York.
(2) Kling, A., Steklis, H.  Brain Behav. Evol., 1976, 13:216-238.

(Acknowledgement: We are indebted to Paul D. MacLean  for providing the original idea for this project.)