According to new research, the most common type of antidepressant can alter the structure of the brain of a developing foetus - if taken by the mother during pregnancy.
Newborn infants exposed to selective serotonin reuptake inhibitors (SSRIs) had greater grey matter volume and white matter structural connectivity in the amygdala and insula compared to infants with no such exposure. These are regions of the brain associated with emotional processing.
The result is concerning, say the researchers behind this study, from Columbia University Vagelos College of Physicians and Surgeons.
That's where SSRI drugs - which help regulate the reuptake of the mood-stabilising neurotransmitter serotonin - can be indispensable. However, the research on the effect of SSRI drugs during pregnancy have given contradictory and inconclusive results.
But now scientists have started to build a picture showing that the effects may be long- rather than short-term.
"Our mouse studies, dating back to 2004, show that inhibiting the uptake of the neurotransmitter serotonin - which is what SSRIs do - during pregnancy has profound effects on foetal brain circuitry," says psychologist and senior study author Jay Gingrich.
"But we don't see behavioural changes in these mice right away. In mice, it looks like SSRIs set the stage for increased anxiety and depressive-like behaviours that emerge later in adolescence."
Although studies using animal models can't always be replicated in humans, there's evidence to suggest that SSRIs do have long-term effects on human development when the foetus is exposed in utero.
A 2016 study, in which Gingrich also participated, analysed the health records of over 15,000 people in Finland, and found that Finnish teens were three to four times more likely to develop depression in early adolescence if they'd been exposed to SSRIs in the womb.
In the new study, the researchers sought to find out what was happening in the brains of infants whose mothers had taken SSRI antidepressants.
They included 98 newborn babies. 16 of those babies' mothers had been treated for depression using SSRIs while they were pregnant; 21 mothers had untreated depression; and the remaining 61 mothers had no history of depression, and were not using SSRIs.
All the babies were physically healthy. The difference could only be seen when their brains were scanned in an MRI machine.
Compared to both the group with untreated depression and the healthy control group, the group of babies exposed to SSRIs showed a significant increase in grey matter volume and white matter connectivity in the right amygdala and the right insula.
This suggests that SSRI exposure has an effect on the brain's sensitivity to serotonin.
It's not quite clear what these results mean, though. Although a higher volume of grey matter in the amygdala and insula has been linked with anxiety, a lower volume has been linked with depression and anxiety too.
It's entirely possible that both increased and reduced grey matter volume can have an effect on mental health. But further study is needed, the researchers said.
"Hopefully these results highlight the fact that something could be going on here," lead author and child psychologist Claudia Lugo-Candelas told Time.
"They point to the fact that there is a signal - we don't know what it means, or don't know how long it might last. But we know it's worth studying."
Meanwhile, the authors don't suggest pregnant women stop taking SSRIs. For one, based on the mouse studies, there's much less risk in the first and second trimesters, so SSRIs may be safe for the majority of the pregnancy.
There may be no "right" answer on what to do.
"It's a difficult clinical decision," said senior author Jonathan Posner.
"We know that maternal depression in and of itself can impact the health of the fetus and the relationship between mother and infant. So, doing nothing is not necessarily the answer.
"But other interventions, such as non-SSRI antidepressants and psychotherapy, can help depressed mums get through pregnancy."
The research has been published in the journal JAMA Pediatrics.