The Assessment of Behavior Symptoms in PPA
According to established guidelines on the diagnosis and management of disorders associated with dementia, the identification of neuropsychiatric symptoms is essential for both the diagnosis and treatment of these clinical conditions as they may constitute the core or supportive diagnostic features of some non-AD dementias [44 ] . In order to detect these behavior modifications, the diagnostic procedures should include a careful medical history obtained from the patient and a close relative, i.e., a person that knows the patient well enough to detect possible changes in his personality.
With respect to PPA, and although the clinical history focuses on the onset, type, and severity of the language impairment, information on the presence of concurrent behavior changes at onset and over the course of the illness should also be gathered. In fact, the reference to significant behavior changes early in the disease course may direct the clinician toward a probable FTLD spectrum disorder rather than AD. At the same time, the clinician should ascertain whether the reported behavior changes emerged prior to language deficits (hence suggesting a more global FTD diagnosis with later language involvement) or if they appeared after the emergence of the language disturbance, which is more likely to reflect the evolution of a PPA syndrome due to FTLD. In any case, both situations should lead to additional biomarker investigation.
Apart from the clinical history, a thorough and accurate detection and quantification of behavior and neuropsychiatric changes should also take into account the use of standard scales. These tools consist mainly of a semi-structured interview with the caregiver. Several behavior scales are available. Specifically in the context of PPA, a recent review of the phenomenology and frequency of neuropsychiatric symptoms in PPA reported that the Neuropsychiatric Inventory (NPI) was the scale used in the majority of studies (67 %) . In fact, the NPI  assesses in detail several behavioral domains, subcategorized as behavioral/comportmental (aberrant motor behavior, disinhibition, apathy/indifference), appetite/eating disorder, mood symptoms (anxiety, euphoria/elation, irritability/lability, depression), and disrup- tive/psychotic symptoms (delusions, hallucinations, agitation/aggression, nighttime behavior). A brief questionnaire form of the scale (NPI-Q)  has been developed with the intention of being used in routine clinical practice and has been shown to be a reliable and valid scale for testing neuropsychiatric/behavior symptoms.
The Frontal Behavioral Inventory [48, 49], originally created to quantify severity of the behavioral disorder in FTD and to allow for the differential diagnosis with AD, is also another scale used in some PPA studies (about 25 %) . Additional scales include the Cambridge Behavioral Inventory (CBI) , which is an informant-based questionnaire comprising 81 items aimed at assessing behavior changes across a range of neurodegenerative disorders. It was originally designed to capture cognitive, behavior, and affective symptoms as well as activities of daily living and evaluates 13 functional/behavior domains: memory, orientation and attention, everyday skills, self-care, mood, challenging behavior, disinhibition, eating habits, sleep, stereotypic and motor behavior, motivation, insight, and awareness. A reduced, 45-item questionnaire has been later developed , and its use has been generalized to major multicentric studies in FTLD . In our own study, we used BDRS to assess behavior changes. The BDRS is a well- validated instrument designed to monitor the progression to dementia [53, 54]. It is easy to administer, provides replicable results, and has high reliability and validity, with scores correlating well with the cerebral changes of primary degenerative dementia . In addition, it has been considered a key instrument to assess functionality in activities of daily living and behavior in AD , being a sensitive and specific screening test for dementia with a good correlation with the neuropsychological test performance . Our results support the use of this simple scale in order to identify, clinically, those patients with more behavior symptoms, despite the major limitation of not assessing several symptoms associated with FTLD (e.g., the occurrence of eating disorders) which are usually addressed by the domain- specific scales mentioned above.