Schizophrenia and Memory

views updated


Schizophrenia affects approximately 1 percent of the population worldwide. It typically involves hallucinations and delusions, also known as psychotic or positive symptoms. Markedly impaired social skills and cognitive deficits, also known as negative symptoms, are also core features. In fact, schizophrenia was originally called dementia praecox (i.e., early onset dementia, to highlight the impairment of cognition. The cognitive domains most often affected by schizophrenia include attention, memory, and language. These deficits are evident even before the onset of positive symptoms (during the so-called "prodrome") and in untreated schizophrenic patients. It is therefore unlikely that the cognitive features of schizophrenia result simply from chronic illness, institutionalization, or medication side effects. Most important, the degree of cognitive decline is the best predictor of functional outcome: that is, how the patient will do in the community once the most severe psychotic features have abated. Here we will focus on abnormalities of memory, describing the type of memory impairment typical of schizophrenia and the cerebral abnormalities that may account for this impairment.

Memory Deficits in Schizophrenia

Despite variable methods and the inherent heterogeneity of schizophrenia, a number of consistent findings appear throughout the literature (Aleman et al., 1999). Here we highlight some of the patterns of memory deficits in schizophrenia:

  • While schizophrenia is associated with a wide range of cognitive difficulties, deficits in memory are particularly pronounced.
  • The memory impairment is selective, affecting primarily explicit memory. Implicit memory, including perceptual priming and procedural memory, appears to be relatively intact.
  • Even within explicit memory, tests of recall ("Tell me everything you saw") are affected more than recognition ("Did you see this before?"). Patients with schizophrenia have difficulty forming associations between items during the encoding (study) period, and do not show the normal pattern of heightened recall after studying words that have common features (i.e., words that are easily associable). Defective recall may therefore be attributable to both poor organization of related items at encoding, and poor search strategies at retrieval.
  • Further examination of the explicit memory deficit indicates a specific impairment in memory that requires conscious recollection of the study episode (i.e., episodic memory) as compared to simple recognition or familiarity-based memory processes. Patients with schizophrenia are less likely to state that they clearly "remember" an event, more often saying that they simply "know" that the event has occurred. Word frequency modulation, which leads to a greater degree of conscious recollection in healthy adults, has no such effect on patients with schizophrenia. Finally, while healthy subjects use both personally relevant and inter-item associations to recollect words, patients with schizophrenia are less likely to do so, relying instead on bizarre and ineffectual linkages.
  • These deficits are not due solely to poor concentration, lack of motivation, distracting positive symptoms, or a medication side effect.

Patients with schizophrenia often suffer from selective impairment of memory that requires the conscious recollection of the study episode, but they retain implicit (and other automatic) memory. Given the recognized abnormality in creating effectual associations, a primary role of impaired "feature-binding" seems to account for the episodic memory impairment in schizophrenia. That is, patients with schizophrenia do not appear to bind together the multifold features of everyday events and therefore find it hard to remember them in detail. Because the ability to place memories in the detailed context of the initial experience is critical for a continued sense of self and the maintenance of a personal history, abnormalities in this realm may relate to the aberrant conscious experiences of schizophrenia.

Neuroimaging of Memory Function in Schizophrenia

Uncovering the neural basis for these deficits in memory is an area of intensive research. Given the role of the hippocampus and the prefrontal cortex (PFC) in normal memory, these regions are the primary focus of this pursuit. Scientists are using both structural neuroimaging techniques (e.g., magnetic resonance imaging [MRI]) and functional neuroimaging techniques (e.g., positron emission tomography [PET] and functional magnetic resonance imaging [fMRI]). In the following section we will briefly summarize the major findings as they apply to schizophrenia (Weiss and Heckers, 2001).

Structural Neuroimaging Studies

The spatial resolution afforded by MRI allows investigators to correlate performance on cognitive tasks with the volume of specific brain regions, including the hippocampus and subdivisions of the PFC. Some evidence suggests just such a link in schizophrenia, with memory performance correlating with the integrity of these two regions. Overall, however, it has been difficult to link memory performance to cerebral structure in schizophrenia, perhaps because the structural changes associated with schizophrenia are subtle. Indeed, the neuropathology of schizophrenia may not involve overt neuronal loss or marked tissue atrophy but rather decreases in the number of only a subset of neurons or deficient connections. These subtle changes may lead to disturbances of cognitive performance without gross morphologic abnormality and therefore would be undetectable with structural neuroimaging. To evaluate the possibility of a functional abnormality within the frontal lobes, the hippocampus, or both, researchers must use methods that allow visualization of cerebral activity.

Functional Neuroimaging Studies

Functional neuroimaging proceeds from the premise that focal neuronal activity relates either to regional cerebral blood flow (rCBF) or to the local degree of glucose metabolism. There are two major classes of functional imaging paradigms: those that examine the resting pattern of activity in a target population (e.g., schizophrenic patients with poor memory) and those that assess changes in cerebral activity during the performance of a task. There are several studies of baseline cerebral activity during memory performance in schizophrenia, but they have yielded discrepant results. Hence, the exact relation between resting cerebral activity and memory performance in schizophrenia remains unclear. Because measures of baseline cerebral activity may not adequately represent the fleeting changes of neuronal activity associated with mental processes, research has shifted to the study of cerebral activity during the performance of memory tasks.

These "cognitive activation" studies attempt to capture the "neural signature" of memory processing—to identify those structures involved at the time of task performance. Researchers obtain images during the performance of a memory task (encoding or retrieval) and during a baseline state. The subtraction of baseline activity from that demonstrated during memory processing allows researchers to infer which regions are "activated" during the memory. Thus it is the change in activity associated with task performance, sometimes called "recruitment," that is commonly reported. Using this type of neuroimaging paradigm, schizophrenic subjects (when compared to controls) appear to show a relative lack of task-related activation in regions known to play a role in normal memory (e.g., PFC and hippocampus).

Frontal Lobe Abnormalities

Patients with schizophrenia commonly demonstrate diminished frontal lobe activation (particularly in the left hemisphere) during memory task performance. Fletcher and colleagues (1998) found that schizophrenic subjects recalled brief word lists (fewer than four words) as well as controls but that they showed a dramatic drop-off in performance with longer lists, a result that suggests a deficit in explicit memory. As word-list lengths increased, both groups initially showed increasing left PFC activation in response. With longer word lists, however, the patients were unable to sustain the degree of prefrontal activation displayed by the normal subjects. These results indicate that schizophrenic subjects are unable to maintain necessary levels of PFC activation in response to increasing task demands.

The relative lack of memory-task-related frontal activation is not a universal finding. Some investigators have demonstrated task-related hyperactivity of the DLPFC in schizophrenic subjects, possibly indicating impaired efficiency of DLPFC function during the performance of a memory task. Task-related hypoactivity may depend on disease-specific characteristics, particularly the degree of negative symptoms.

Deficient frontal activity has been linked to "dysexecutive" features of memory; the inability to organize information for optimal performance. While control subjects use the implicit semantic associations between words to facilitate encoding and support subsequent retrieval, schizophrenic subjects do not use this strategy. Instead, patients tend to use ineffectual or bizarre associations, leading to poorer recall. The inability to use the "normal" associative mechanisms that take advantage of inherent properties of the items to be remembered correlates with a lack of activation in the inferior frontal lobes bilaterally. Difficulty in organizing encoding and retrieval strategies in schizophrenia therefore appear to be associated with lower memory task-related metabolism in regions of prefrontal cortex.

Temporal-Lobe Abnormalities

Recent studies in patients with schizophrenia demonstrate abnormal baselin levels of activity in the temporal lobe and memory task-related hypoactivity of the hippocampus. Heckers and colleagues (1998) studied subjects after they learned words using a deep semantic encoding strategy or a shallow perceptual one. During scanning, subjects were asked to complete three-letter word stems of the target words. Normal subjects recalled more words after deep semantic encoding, which was associated with a recruitment of the hippocampus; the attempt to retrieve the words encoded using the shallow perceptual strategy resulted in bilateral activation of the prefrontal cortex. Schizophrenic subjects recalled fewer words after deep semantic encoding, which was associated with an impaired recruitment of the hippocampus. Increased hippocampal activity at baseline and impaired recruitment during episodic memory retrieval might represent the functional correlate of an abnormal cortico-hippocampal interaction in schizophrenia. These findings complement the studies of abnormal frontal cortex activity in schizophrenia. A frontal-hippocampal disconnection in schizophrenia may involve temporal-lobe circuits that play a role in the processing of language, pathways to and from the hippocampal formation, or both.

As with the findings of frontal hypoactivity, abnormalities in temporal lobe activation are not universal. Several studies, despite showing impairments of explicit memory in the schizophrenic cohort, have failed to demonstrate functional abnormalities in the temporal lobe. It is unclear whether this is due to differences in experimental design, image acquisition, disease characteristics, or other as yet unknown factors.


The results reviewed here suggest an abnormality of hippocampal and prefrontal cortex function during memory in schizophrenia. The pace of discovery in this area has been quite rapid. The next major strides in the understanding of the neural basis of schizophrenia await the advent of even better technology that would allow more precise temporal and spatial assessment of neural functions.



Aleman, A., Hijman R., de Haan, E. H. F., and Kahn, R. S. (1999). Memory impairment in schizophrenia: A meta-analysis. American Journal of Psychiatry 156, 1,358-1,366.

Fletcher, P. C., McKenna, P. J., Frith, C. D., Grasby, P. M., Friston, K. J., and Dolan, R. J. (1998). Brain activations in schizophrenia during a graded memory task studied with functional neuroimaging. Archives of General Psychiatry 55, 1,001-1,008.

Heckers, S. (2001). Neuroimaging studies of the hippocampus in schizophrenia. Hippocampus 11, 520-528.

Heckers, S., Rauch, S. L., Goff, D., Savage, C. R., Schacter, D. L., Fischman, A. J., and Alpert, N. M. (1998). Impaired recruitment of the hippocampus during conscious recollection in schizophrenia. Nature Neuroscience 1, 1,318-1,323.

Weiss, A. P., and Heckers, S. (2001). Neuroimaging of declarative memory in schizophrenia. Scandinavian Journal of Psychology 42, 239-250.


Anthony P.Weiss