Harriet Holliday sparkled with personality. She reminded Kevin Horowitz, her third husband, of the mother in Mamma Mia!—free-spirited, flamboyant and nurturing. She dressed with a sexy, sophisticated glamour and regaled friends with witty tales at soirees. As hospitality manager at a winery in Napa Valley in California, she planned events for hundreds of guests. But around six years ago, at age 49, Holliday “started turning mean,” Horowitz recalls. “She didn’t know when to hold her tongue.”

She became touchy and sarcastic, alienated friends and was soon fired from her job. Other odd or inappropriate behaviors surfaced. When the couple dined out, Holliday would wear a stylish dress with house slippers. She flirted outrageously with young male strangers and tickled them.

After years spent exploring possible explanations for Holliday’s strange transformation, from menopause to depression, the couple learned in 2009 that she had a little-known, incurable and fatal disease: frontotemporal dementia (FTD), in which significant portions of the brain’s frontal and temporal (side) regions degenerate.

Most of us think of dementia as a thief of memory and thinking, but this particular variety instead steals away social graces, emotions and empathy. Often the first signs are “changes in the way that people relate to others that make them unlikable,” says neurologist Bruce L. Miller, director of the Memory and Aging Center at the University of California, San Francisco, where Holliday’s illness was definitively diagnosed.

Frontotemporal dementia is actually the most common form of dementia in people younger than 60, typically striking earlier than Alzheimer’s disease and culminating in death within roughly eight years. By one estimate, 15 out of 100,000 people between the ages of 45 and 64 develop it. Patients lose the ability to connect with others, but because self-awareness also disappears, they remain ignorant of their deficits.

By revealing what human behavior and personality look like when stripped of emotional connectedness, the tragic disorder starkly illustrates how profoundly our species is wired to be social. Studying people with this social dementia offers a unique opportunity to understand the neural underpinnings of self-awareness, of certain complex emotions that emerge only in social situations and of basic personality traits. “You can’t take a human being and cut part of their brain out and see whether they’re still empathic or whether their personality changes,” says U.C.S.F. neuropsychologist Katherine P. Rankin. But this illness is nature’s equivalent of that experiment.

Studies of patients have, for example, revealed that personality—the pattern of habits and emotional tendencies that distinguishes individuals—does not simply emerge from a soup of brain chemicals but can be traced to specific brain structures and circuits. In particular, personality may reflect how effectively certain brain regions are interconnected. One such network that monitors self-awareness and social and emotional cues—and is ravaged in FTD—seems to play a key role in interpersonal warmth and empathy. Eventually researchers might know enough to determine a person’s capacity for, say, extroversion or anxiousness from scans that assess the size of relevant brain structures and the strength of their connections with one another.

The “Other Dementia”
First described in 1892 by Czech neurologist Arnold Pick, the classic form of FTD (also called Pick’s disease) was tricky to diagnose and largely went overlooked for nearly a century. In the 1980s neurologists in the U.S. believed Alzheimer’s disease was the only significant cause of dementia. But in 1987 Swedish neuropathologist Arne Brun and psychiatrist Lars Gustafson autopsied 158 dementia sufferers and discovered that 13 percent of them had Pick’s or other types of FTD.

Miller was among a handful of researchers who in the 1990s characterized the peculiar social decline that typically occurs in FTD and established it as a separate entity from Alzheimer’s. People with Alzheimer’s retain their social skills in the early stages, and many remain warm and sensitive to the end. Initially their illness mainly devastates the back-of-the-brain regions involved in memory, language and visuospatial skills, only later spreading toward frontal areas. FTD often spares those posterior brain areas [see bottom illustration on opposite page]; its biggest symptoms are, instead, disinhibition leading to disruptive behavior, a shift toward emotional coldness, and apathy. Patients also often overeat and lose their judgment, sometimes losing control over their finances and going bankrupt.

Miller believes that many cases are still overlooked, because spouses and family—and even many medical professionals—usually view the first changes as signs of a midlife or marital crisis. Of the frontotemporal patients who come to U.C.S.F., roughly half are initially misdiagnosed, most commonly with Alzheimer’s, depression, bipolar disorder or, occasionally, schizophrenia.

No treatment exists yet for FTD, although doctors may try to temper behavioral troubles with psychiatric drugs and psychologists can offer supportive guidance. Researchers know that in about half of patients, a protein called tau (also implicated in Alzheimer’s) forms in toxic clumps inside neurons in frontotemporal areas. In most other cases, the brain cells are sickened from pileups of a protein called TDP-43, which is also a culprit in amyotrophic lateral sclerosis, or Lou Gehrig’s disease.

Ground zero of FTD seems to be an unusually large, spindle-shaped kind of neuron that exists only in the brain’s frontal lobes—specifically, in the frontal insula and anterior cingulate cortex. In a 2006 study U.C.S.F. neurologist William W. Seeley and his colleagues analyzed brain tissue from deceased FTD patients and observed that in early stages, the disease selectively decimates these cells, called von Economo neurons (VENs). Brain-scan studies suggest that the frontal insula and anterior cingulate are active when we are hungry, thirsty, in pain or see another person in pain. The insula, in particular, monitors both our own bodily sensations as well as our gut feelings about others. Thus, the two regions may be important in sizing up emotionally charged social interactions, and, the researchers speculate, the large size of the VENs may enable them to quickly communicate the gist of those calculations across the brain. When death strikes VENs and the neural network they belong to, “patients lose their social humanity,” Seeley says. “It slowly fades away.”

No Inhibitions
In 2008, after learning that Holliday had run up a shocking $74,000 on credit cards, her husband finally realized something was seriously wrong. Although one local physician pronounced an Alzheimer’s diagnosis, another specialist suspected FTD and referred her to U.C.S.F. Every year the Memory and Aging Center, one of the world’s largest FTD clinics, sees roughly 100 patients with social dementia as part of a multimillion-dollar research program funded by the National Institute on Aging.

I met the couple last June when they went to U.C.S.F. for Holliday’s second annual assessment—a four-day battery of MRI tests and evaluations by neuropsychologists who probed her cognitive, social and emotional status with verbal questions and pencil-and-paper tests. Now age 55, she was not the angry woman of six years ago: she seemed docile and quiet but removed, with little expressiveness in her face.

With antidepressant treatment, her cranky behavior had faded into blandness, but she had also grown apathetic. “She doesn’t have a lot of the emotions the rest of us have,” Horowitz, a man with worried eyes in a sad, worn face, told me. At a funeral they attended, she did not get what the fuss was about. The old Harriet was a fashion maven; now her husband had to lay out her clothing (khakis, button-down shirts) every morning, or she would wear yesterday’s outfit. Her impulse control was gone; recently, at the gym, she followed a guy into the men’s locker room. Horowitz enrolled her at an adult day care center three days a week. Like a three-year-old, she needed constant supervision.

A few weeks later, in July, Miller convened with half a dozen colleagues to review Holliday’s case. On the standard so-called Mini-Mental State Examination of cognitive health, she scored within normal, but results on other tests highlighted some severe deficits in problem solving and other executive functions, cognitive losses that are common in moderate to advanced stages of the disease. And during testing, she was inappropriately goofy and giggly, at one point imitating squirrel noises.

Miller turned to a laptop to scroll through black-and-white MRI images of Holliday’s brain. Strikingly, her frontal lobe borders were edged with an encroaching blackness, reflecting tissue loss, instead of being nicely filled out with light-colored areas of healthy wrinkled tissue. Other brain regions toward the back—“Alzheimer’s disease central”—were largely intact. “You’d have to be almost an idiot not to be able to diagnose this,” he said.

The widespread frontal damage included atrophy in the insula, anterior cingulate cortex and another key target, the orbitofrontal cortex, the area behind and above the eyes that helps in decision making by sensing potential reward or punishment. All three regions normally cooperate with the amygdala to produce complex feelings and social behavior. For instance, structural imaging studies by U.C.S.F. neurologist Howard Rosen correlated orbitofrontal loss with disinhibition and anterior cingulate damage with apathy. Meanwhile, on the right side of Holliday’s brain, her anterior temporal lobe, which is important for recognizing emotion and faces, was “paper thin,” Miller said.

And yet, Miller said, her behavior was better than that of many patients. One man dropped his pants when he was having his blood drawn. Another, a retired surgeon, broke into his neighbor’s house to steal liquor and made sexual advances toward women at a wedding rehearsal dinner. About half of FTD patients had engaged in antisocial behaviors that could (or did) get them arrested, including shoplifting, running a stoplight or driving while drunk.

Short-Circuiting the Self
To tease out the emotional blind spots in social dementia, U.C.S.F. investigators send patients to the laboratory of clinical psychologist Robert W. Levenson at the University of California, Berkeley. Levenson’s group monitors how people react during tests designed to elicit specific emotional responses by rigging the subjects with sensors that track heart rate, blood pressure, breathing, sweating and other measures, second by second. Facial expressions and body movements are videotaped and coded for analysis.

One morning in his home office, Levenson played me a video file from Holliday’s 2009 emotions assessment. The video showed her hooked up to sensors, staring at a big X displayed on a television screen. Suddenly an explosive bang sounded. She instantly jumped and clutched at her heart, blue eyes wide with fear. Then she just sat there, looking around. Levenson then played clips of healthy test subjects. When the bang went off, one woman let out a loud, shocked gasp, with an exclamation—“God!”—that quickly segued into laughter. This acoustic startle task usually triggers a basic, fearful fight-or-flight reflex, but in the lab, people with intact frontal lobes realize they are being studied, so they also self-consciously chuckle at their reactions. But like Holliday, most social dementia patients show no signs of embarrassment or amusement, Levenson and his colleagues found.

Although basic emotions such as fear, anger, sadness and happiness still function at a low level in FTD, the more complicated social or self-conscious emotions, including embarrassment, guilt and shame, take a big hit. Those feelings require self-awareness and the ability to gauge how your behavior stacks up against social norms—mental processes that FTD patients lack.

In addition to their lack of self-consciousness, patients can also be shockingly callous. When shown scenes from movies, for instance, they typically have trouble recognizing when a character is embarrassed or ashamed and evince little distress when shown images of human suffering. They often display a disheartening loss of compassion in real life, too. One day a woman I will call Alicia briefly blacked out at home while her husband, a retired travel agent who had FTD, was out. Alicia called her son to come over. The husband, a gardener, later drove up with a flat of seedlings. When the son explained that they needed to take Mom to the hospital, his father replied, “I have to first take care of my seedlings,” Alicia recalls. “And this is a man that I’m married to, at that point, 44 years,” she says with bitterness. He died last May.

From the caregiver’s viewpoint, Levenson says, the social dementia has revealed another lesson. Seeing someone lose his or her emotions is even tougher than watching a person’s cognition crumble. “It turns out this is much worse, or it’s more hurtful, when someone you love still knows who you are but no longer cares about you,” he adds. That rejection does not make you sad. “It makes you furious.”

Hunting for Personality in the Brain
Although the neuroscience of sentiment is highly complicated, an even bigger conundrum is how the cerebral cortex might blend emotions with behaviors and thoughts to form personality. Only in the past few years have researchers used functional MRI to map out brain circuits for the so-called big five dimensions of personality: openness to new experiences, conscientiousness, extroversion, agreeableness and neuroticism. A few investigations have, for example, traced extroversion to the orbitofrontal cortex.

Though invaluable, fMRI findings from normal brains cannot provide a complete picture for any personality trait or social behavior, Rankin says. For instance, when a dozen neural areas activate during empathy—an emotional response involved in extroversion and agreeableness—that result does not tell you which areas are essential for this feeling, as opposed to just playing a supporting role. On the other hand, Rankin’s structural MRI analyses of FTD patients with known brain lesions have shown that real-life empathic abilities are lost when the disease has eaten away specific sections of the temporal lobe—including its forwardmost section, the temporal pole—as well as patches of the orbitofrontal cortex, all on the brain’s right side.

Rankin and other U.C.S.F. investigators have been combining patients’ structural brain scans with their psychological evaluations in an attempt to link particular brain regions to personality traits. For instance, in a study published in November 2009 she and neurologist Marc Sollberger asked family caregivers of 214 people with neurodegenerative diseases, including FTD, Alzheimer’s and mild cognitive impairment, to rate the patients on 64 adjectives from “shy” to “iron-hearted.” Final scores yielded an overall assessment on four pairs of interpersonal traits related to the socially relevant dimensions of extroversion and agreeableness: dominance versus submissiveness, coldness versus warmth, introversion versus extroversion, and arrogance versus ingenuousness.

On average, compared with 43 healthy elderly individuals, patients with FTD were less dominant—meaning, less assertive or pushy—and less extroverted, a pattern associated with gray matter loss on MRI scans in certain parts of the frontal lobes, particularly on the brain’s left side. Alzheimer’s patients showed a similar but smaller shift, a sign of disease having spread to some frontal areas. But only the FTD patients showed decreased warmth, a change associated with atrophy in structures in the right hemisphere, including parts of the temporal lobe, the orbitofrontal cortex, frontal insula and amygdala. “Those structures seem to be necessary to be a warm, connected person,” Rankin concludes.

In the simplest terms, social aspects of personality are a balancing act between brain areas that give rise to forceful behavior to get what you want and brain areas that help you understand and care about another person, she says. Holliday, for example, could no longer tell when a voice sounded sad, happy or sarcastic, deficits partly caused by atrophy of the temporal lobe, but she wasn’t as cold or introverted as other patients. Meanwhile the extensive devastation of her frontal lobes produced a shift toward submissiveness and a Forrest Gump ingenuousness. On the whole, Rankin’s results demonstrate that personality is not merely an epiphenomenon of brain chemistry. Instead interpersonal traits “actually have an anatomy to them,” she says.

Given Rankin’s research, it is seductive to think, “Gee, maybe that aloof neighbor who never says hello just has a puny right orbitofrontal cortex.” Of course, in a healthy person, functioning with a relatively small orbitofrontal area may not be equivalent to having one that has been destroyed by disease. But so far Rankin’s findings are fairly consistent with the emerging literature on the neuroscience of normal personality.

Personality Networks
Rankin and others emphasize that personality traits will never be reduced to any single brain structure. Probably more important is how well particular regions interconnect and cooperate. Powerful new insights about brain function in general have been emerging from the discovery of intrinsic connectivity networks (ICNs), neural regions whose activity fluctuates in tight synchrony when the brain is idling or unengaged in any mental task.

Some of the same key brain zones that Rankin has linked with warmth toward others, it turns out, are part of an ICN that her colleague Seeley identified and calls the “salience network.” It is anchored by the frontal insula and the anterior cingulate cortex. Seeley and his collaborators at Stanford University theorize that this network rapidly filters through the sensory signals, internal bodily sensations, social cues and emotions that deluge the brain every second, then zeroes in on the most pressing problem at hand. It singles out what is important or salient, whether the issue is hunger or a tense social dilemma.

But when this salience network breaks down in individuals with FTD, Seeley says, they cannot pick up social signals or feel the weight of their next actions. Put another way, patients stop caring about loved ones because they “no longer realize that other people are important,” Rankin notes. Additional networks most likely play a role in other aspects of warmth and empathy. For instance, Seeley has identified another potential ICN underlying emotional empathy that is anchored by the temporal pole and also disrupted in one type of FTD.

Seeley hopes to refine methods for detecting differences in how “functionally” connected, or synchronized, brain regions within ICNs are in frontotemporal patients; such tools could help diagnose the disorder earlier or track responses to experimental treatments. He and Rankin are also studying whether these functional connectivity levels match up with patients’ social behavioral ratings. The researchers think we are headed toward a day when brain-scan analyses of the size and connectivity of particular neural structures will grow precise enough to gauge a person’s capacity for empathy or to predict some very basic aspects of personality, such as arrogance or anxiousness. Such technology could have interesting applications.

“You can imagine a dating service that requires people to come down to its office and get in the scanner for 10 minutes” as part of their personality assessment, Seeley says, chuckling. Alternatively, if a brain MRI of a patient receiving counseling revealed a strong response in areas linked with compassion, such a result might greatly inform therapy, Rankin says. On the other hand, weak connectivity in a region important for warmth could point to a different therapeutic target. Some scientists, however, are skeptical that insights into the anatomy of personality will ever be advanced enough to enable such tantalizing uses.

The Road Ahead
Meanwhile researchers at U.C.S.F. and elsewhere are testing a few treatments for social dementia. In early 2011, for example, investigators at the San Francisco clinic and other centers plan to start a trial with an experimental Alzheimer’s drug called Rember in FTD patients. Made by Singapore-based TauRx Pharmaceuticals, the drug is designed to block toxic tau protein buildup inside neurons. Holliday could be eligible for the experimental  treatment, but with her substantial brain atrophy, it is unlikely to help much.

The illness has shredded Holliday and Horowitz’s plans to retire to Mexico. Horowitz, who owns a construction firm in Napa Valley, now foresees a far bleaker journey. “I’m walking next to her and going all the way down to hell,” he told me last summer. And yet, amid this nightmare, he says he is determined “to find something in it that is valuable to someone.” That mission explains why they participate in the U.C.S.F. research program. The disintegrating brains of Holliday and others are revealing clues to a cure—and to what makes us human.