Dream reports may aid in diagnosing mental diseases

Brazilian researchers are developing a technique to mathematically analyze dream reports to help identify symptoms of schizophrenia and bipolar disorder.

By Elton Alisson

Agência FAPESP – The notion presented by Sigmund Freud (1856-1939) in the 1899 book “The Interpretation of Dreams,” that “dreams are the royal road to the unconscious,” was key to psychoanalysis and may also be useful in psychiatry, specifically in the clinical diagnosis of mental disorders such as schizophrenia and bipolar disorder.

This assessment was made by a group of researchers at the Brain Institute of the Federal University of Rio Grande do Norte (UFRN), in collaboration with colleagues from the Department of Physics at the Federal University of Pernambuco (UFPE) and the Research, Innovation and Dissemination Center for Neuromathematics (Neuromat) – one of the FAPESP Research, Innovation and Dissemination Centers.

These researchers have developed a technique that can be used to mathematically analyze dream reports. In the future, this technique could help diagnose psychoses.

The technique is described in an article published in January in Scientific Reports, an open access journal of the Nature Publishing Group.

“The idea is that a relatively simple and inexpensive technique could be used as a tool to help psychiatrists make more precise clinical diagnoses of patients with mental disturbances,” Mauro Copelli, UFPE professor and one of the study’s authors, told Agência FAPESP.

Copelli – whose master’s and doctoral degrees were partially funded by scholarships from FAPESP – has stated that despite considerable efforts to increase the precision with which mental disorders are classified, the current method used to diagnose psychoses has been subject to harsh criticism.

This criticism arises because the method still lacks objectivity and because most mental disturbances have no biomarkers (biometric indicators) that can help psychiatrists diagnose them more accurately.

Furthermore, patients with schizophrenia or bipolar disorder often present common psychotic symptoms such as hallucinations, delirium, hyperactivity and aggressive behavior, which may confound diagnosis.

“Diagnosing psychotic symptoms is highly subjective,” Copelli stated. “That is why the most recent version of the Diagnostic and Statistical Manual of Mental Disorders [published by the American Psychiatric Association in 2013] came under so much fire,” he said.

To develop a quantitative method that can be used to evaluate psychiatric symptoms, the researchers, with the consent of those involved, recorded the dream reports of 60 volunteer patients treated at the outpatient psychiatric unit of a public hospital in the city of Natal (RN).

Some of the patients had already received a diagnosis of schizophrenia or bipolar disorder. The remaining patients, who made up the control group, presented no symptoms of mental disorders.

The patients’ dream reports were provided to psychiatrist Natália Bezerra Mota, a doctoral candidate at UFRN and first author of the study, and were subsequently transcribed.

Using a software program developed by Brain Institute researchers, the patients’ speech samples were converted into graphs. These graphs are mathematical structures similar to diagrams, in which each word uttered by the patient was represented by a point or a node, as in a line of crochet.

In analyzing the graphs from the dream reports provided by the three groups of patients, the researchers noted that there were clear differences among them.

The size, in terms of the number of loops or links, and the connectivity (relationship) between the nodes of the graphs of patients diagnosed with schizophrenia, bipolar disorder and those without mental disorders presented variations, said the researchers.

“Patients with schizophrenia, for example, give reports whose graphical representations have fewer connections than those of the other groups of patients,” Mota said.

Speech differences

According to the researchers, the patients could be differentiated on the basis of graphs of their dream reports because their speech characteristics are also quite different.

Schizophrenic patients often speak laconically with little digression (subject departure), which explains why there is less connectivity and fewer loops in the graphs of their dream reports than those of bipolar subjects.

Patients with bipolar disorder tend to present the opposite symptom to digression, referred to as logorrhea or excessive verbiage, which involves talking in meaningless sentences in a confused way. This symptom is known in psychiatry as a “flight of ideas.”

“Through graphical analyses, we found that these measurements were significantly correlated with negative cognitive symptoms measured by psychometric scales used in the clinical practice of psychiatry,” Mota said.

By converting the striking aspects of the patients’ speech into graphs, it is possible to develop a computational classifier that can help psychiatrists diagnose mental disorders, stated Copelli.

“All the speech occurrences in patients with mental disorders that have an apparently geometric significance in the graph can be mathematically quantified. These data can then be used to help categorize a patient as schizophrenic or bipolar, with a success rate that is comparable or even better than the subjective psychiatric scales used for this purpose,” he said.

The researchers’ next goal is to evaluate a larger number of patients and calibrate the algorithm (sequence of commands) of the software that converts the dream reports into graphs, with the aim of using the software for large-scale use in clinical psychiatry.

In addition to its initial use in the diagnosis of psychoses, the technique could be expanded to include several other uses, said Mota.

“It may be used, for example, to seek additional information about the language structure that can be applied in analyzing the reports of individuals who not only have psychotic symptoms but are also in various stages of cognitive decline, such as dementia. On the other end of the spectrum, it can be applied during the process of learning and developing speech and writing,” the researcher said.

Role of dreams

During the study, the researchers also developed and analyzed the graphs of reports about activities carried out by the patient volunteers immediately before the dream.

The graphs of these day-to-day reports, referred to as “waking reports,” were not as indicative of the type of mental disorder suffered by the patient as were the other graphs, Copelli said.

“We were able to distinguish schizophrenics from the rest of the groups by analyzing graphs from the waking reports, but we were not able to distinguish the bipolar and control group subjects very well,” Copelli said.

The researchers still do not know why the speech graphs derived from the dream reports provide more information about psychosis than the graphs derived from the waking reports.

Some hypotheses examined in Mota’s doctoral research are related to the physiological mechanisms involved in memory formation.

“We believe that, because dreams are temporary memories, they may be more cognitively demanding and have greater affective impact than memories related to daily life, and this may make the reports more complex,” said the researcher.

“Another hypothesis is that dreams are related to events experienced by only one person and not shared with anyone else, which is why they may be harder to explain than an activity related to everyday experiences,” she said.

To test these hypotheses, the researchers plan to expand their data collection by conducting surveys of patients who have reported their first psychotic episode to clarify whether other types of reports, such as those of old memories, can be matched to dreams in terms of psychiatric information. They also want to establish whether they can use the method to identify signs or sets of symptoms (prodromes) and to monitor the effects of medications.

“We plan to conduct laboratory experiments using high-density electroencephalography and various techniques for measuring semantic distances as well as graphic structure analysis to determine how the stimuli received just before sleeping influence the dream reports produced upon awakening,” said Sidarta Ribeiro, researcher at the Brain Institute at UFRN.

“We’re particularly interested in the different effects of images with affective value,” said Ribeiro, who is also an associate researcher at Neuromat.

The article “Graph analysis of dream reports is especially informative about psychosis” (doi: 10.1038/srep03691), by Mota and others, can be accessed at the Scientific Reports website at www.nature.com/srep/2014/140115/srep03691/full/srep03691.html.