There have been intimations of this. 10 years ago [ Proc. Natl. Acad. Sci. vol. 100 pp. 11702 – 11705 ’13 ] Full term infants were shown to respond more to human speech played forwards than when the tape was reversed. A clever technique called optical topography was used — it is noninvasive, and relies on the thinness of the neonatal skull.
Now comes [ Proc. Natl. Acad. Sci. vol. 110 pp. 15145 – 15150 ’13 ] The authors presented variants of words to fetuses; unlike infants with no exposure to these stimuli, the exposed fetuses showed enhanced brain activity (mismatch responses) in response to pitch changes for the trained variants after birth. (What this means is that there is (again a noninvasive) way of measuring brain activity — and there is greater activity when an unexpected variant is presented. We are novelty seekers from the get go.
Also there was a significant correlation between the amount of prenatal exposure and brain activity, with greater activity being associated with a higher amount of prenatal speech exposure. Even more impressive, the learning effect was generalized to other types of similar speech sounds not included in the training material.
We know know that the infant brain gets tuned to the language they hear. Japanese infants can hear the r sound at birth (again brain electrical activity responds to it) but after 6 – 9 months of listening to a language without the sound, their brain becomes tuned so they can’t.
So maybe pregnant ladies should listen to Mozart