When it comes to predicting whether someone will have Alzheimer’s disease, newfangled diagnostic tests for the illness aren’t as good as old-fashioned quizzes of thinking and memory.

That’s the word from a study that compared different methods for identifying Alzheimer’s. the results was just published in the Archives of General Psychiatry.

Scientists have been looking for a scan or lab test, such as changes in particular proteins in the blood, to simplify and speed up the detection of Alzheimer’s. the goal is to be able to identify people at risk of the illness long before symptoms such as impaired thinking surface. But that goal is far from being realized.

In this study, researchers compared how well biological and cognitive tests did at identifying whether people who already had mild cognitive impairment would go on to develop Alzheimer’s. they found that two tests measuring verbal memory were the best at predicting whether someone would develop Alzheimer’s in the next two years.

In those tests, people were read a story or a list of words, and then asked to recall the information 30 minutes later. In people who are progressing to Alzheimer’s, “there’s forgetting,” neuroscientist Terry Goldberg, a co-author of the study, told Shots. “There’s an increased rate of forgetting.”

One biological marker, a brain scan that measured changes in the thickness of the middle temporal lobe, also predicted people who went on to have Alzheimer’s. Other biological markers, such as tests for telltale proteins in blood and cerebrospinal fluid, also revealed changes, but not as well. Those tests have been used in research laboratories, but aren’t being used to test for Alzheimer’s in clinical practice.

“It’s not that the biomarkers didn’t work,” says Goldberg, a professor at Hofstra North Shore-LIJ School of Medicine in Manhasset, N.Y. “But when you compared them head to head, you find that the cognitive markers are the most robust.”

That means that the thinking tests can also be used for guides in designing therapies. In fact, Goldberg says, they may ultimately be more useful that the biological tests. “What you want to change is how people do in everyday life,” he says.

The trick is to diagnose mild cognitive impairment, a stage between normal forgetfulness and serious cognitive problems. It sometimes leads to Alzheimer’s, but can be caused by other diseases, too. People with MCI can be more forgetful, have trouble multitasking, and take longer to remember or decide. this study looked at 116 people with MCI who were diagnosed with Alzheimer’s disease in two years, 204 people with MCI who didn’t get Alzheimer’s and 197 healthy people as controls.

The study used information from the Alzheimer’s Disease NeuroImaging Initiative, which is looking for ways to identify risk factors for Alzheimer’s, and techniques for early diagnosis.

The goal would then be to create treatments that could profoundly slow or even halt the disease’s process. No such treatments exist now. Earlier this year, the National Institute on Aging redefined Alzheimer’s, broadening the definition and including the fact that biomarkers can be used for early detection. But since there is no treatment that slows or halts the disease, those biomarker tests don’t help patients or families manage the disease.