What goes through the mind of a driver in the crucial moments before a fatal car collision? Are they trying to avoid that tree or telephone pole—or are they steering right for it?

A new study looks into a crash involving a mother driving her four children to school—and concludes that familial murder-suicide by car could be an underreported phenomenon not identified by investigators at crash scene because intent may be impossible to positively prove.

“While certain events such as single-vehicle crashes into trees or other objects are well recognized as being potential occult suicides, the possibility of multiple deaths in a family in a crash being due to a familial murder-suicide is less well appreciated,” writes the Australian team in the Journal of Forensic Sciences.

The case involved a 39-year-old mother driving her four children to school one morning: a 14-year-old girl, an 11-year-old girl, a 9-year-old boy and a 6-year-old boy.

Normally at a main road turnoff, she would have dropped one of her daughters off to wait for her bus, as she had done many times before.

But this time, she kept driving a short distance down the same road, before stopping and turning around, according to the paper.

One of the surviving kids in the car later reported that she then drove at an unusually high speed, missing a turnoff—and colliding straight into a tree, the study recounts.

The mother was dead at the scene, with skull fractures and a broken face, cerebral contusions and a subarachnoid hemorrhage, among many other injuries. She was not wearing her seat belt at the moment of impact.

The teenaged daughter died five days later from loss of oxygen to her brain, due to major trauma to the head, among a litany of other injuries. She had been wearing her seat belt, however.

The three other kids were all hospitalized with serious injuries that were non-lethal.

The conditions for driving were not hazardous at the time of the crash. Follow-up investigation revealed that the mother had “recent depressive symptoms with paranoid ideation,” according to the researchers, from the University of Adelaide in Australia.

But still, detectives could make no conclusive determination.

“Police investigators attributed the crash to driver error but also could not exclude the possibility of a nonaccidental etiology,” according to the paper.

Murder-suicides are rare at a population level—but occur steadily each year. Between 1,000 and 1,500 are reported each year—and they are most often in the family.

Motivations generally depend on the relationship: spouses kill their partners over jealousy or revenge, or due to financial or social pressures. The majority—55 percent—are married couples.

Parents kill children due to perceived “mercy killing,” or out of “altruistic” feelings, according to the paper. Mothers will most often use suffocation and poisoning to kill their kids, while fathers will generally use strangulation, guns and knives while acting as so-called “family annihilators.”

The case in question showed how this case could be a murder-suicide that just could not be proven, based on the facts available to police, the authors conclude.

“While this could have been a simple crash, there was no indication at the scene of dangerous driving conditions,” they write. “In addition, unusual behavior such as electing to drive the children to their school bus but not dropping them off with speeding may suggest that there was a nonaccidental component.”

The same group of Australian researchers has investigated the phenomenon of single car, driver-only suicides in that country, concluding in 2012 that most of the deceased were young and under the influence of alcohol at the time of impact.

Another team recently looked at murder-suicides by car in which victims were chosen at random—and found that generally suicides “are able to act with great consideration for the consequences of their actions.”