M. Maranto, 
psycologist 

G. Porzionato
professor of general psycology and personality

The autors conduced an investigation above all senile disease. Their experience lead them to believe that the music had beneficial effects related to the perception of well-being expressed by mood. It seems that the mnemonic, emotional and cognitive activations that music stirs played a fundamental role.
They do not intend to claim that music therapy is the only factor responsible for the lower mortality rate, but to acknowledge that an emotional variation capable of positively stimulating the perception of well-being of an individual can also affect maintenance of their cognitive abilities and general physical condition.

Introduction
How does music alter the psychophysical state of an elderly person? Can it somehow influence the length of his/her life? Answering these questions is undoubtedly complicated, given the complexity of the problem. In an intervention at the "Music Therapy and Mental Disorders" conference held in Genoa, Mancia (1989) quoted numerous studies on musical perception, memory and emotions which demonstrated that the human brain elaborates music in a modular way.
LeDoux (1994, 1996) claims that the limbic system is responsible for the emotional connotation of a memory and assigns the task of elaborating its cognitive part to the cortex. Taking into account, however, that the cortex, hippocampus and thalamus are directly connected to the lateral nuclei of the amygdala, it appears that the final product of memorization comes from an interaction that is spread through numero us cerebral sites and at different levels.
Goleman (1995) and Ciompi (1993) are also of this opinion. The latter puts forward the hypothesis of "affective logic": in this, states of mind, feelings and emotions are defined as complex phenomena that also operate at a vegetative level through hormones. The cognitive-intellectual functions situated at cortical level are closely connected to "their" emotional part via the thalamus and the amygdala. Events with strong emotional connotations are thus memorised taking into consideration the state of mind at that moment, and this will have a strong influence when they are subsequently retrieved.
During the seventies Grof (1975) induced particular feelings through hypnosis or psychogenic drugs which elicited the memories related to these. Scapagnini and Scapagnini (1997) uphold the theory of psycho-neuro - endocrine - immunology (PNEI), highlighting the strong interaction between the nervous system, the immune system and the endocrine system, and the importance of observing the totality of the human body from various points of view and of operating on a global level to improve its efficiency. For his part, Leaf (1974) demonstrated that psychological well-being is a significant factor in subjects who maintain an enviable physical condition while ageing.
In this field musical experience and the recollection of it can play an important role. First of all, musical memories are laden with emotional content which is very easily retrieved in memories even in those who are seriously debilitated from a cerebral point of view. It also seems that the emotional part of memories can at least partly reactivate the more cognitive part and this can be explained by what was previously stated (LeDoux, 1994; Ciompi, 1993). It is also very clear that both emotional and cognitive memories can evoke positive states of mind since they have been memorised in particularly joyous moments, the moments of music. This is also plausible following the previous hypothesis (Grof, 1975).
Finally, it is common sense to believe that the improvement of well-being caused by pleasant memories influences the entire neuro- endocrine- immunologic apparatus in the long term, as Scapagnini and Scapagnini (1997) and Leaf (1974) suggest, thus also contributing to maintaining physical health during old age.

The empirical experience
Listening music therapy was introduced as an experiment in 1995 in an organisation which is made up of 13 old people’s homes in the Veneto region. The music therapy was aimed at the non self-sufficient residents in two three and four-floor buildings and in two five and six-floor "sections".
The residents were there f o re spread over eighteen f loors in all, mixed as regards their pathologies (above all senile or multi-infarct dementia, hemiplegia, vascular disorders, Alzheimer’s and Parkinson’s disease). The music therapy activity was performed weekly on each floor from the 31st May until 30th November 1996.
The objective of this empirical experience was to create a space for musical listening where subjects could socialise, pass a moment of recreation and retrieve emotions and memories through listening to pieces of music.
The study underpinning this intended, however, to perfect an instrument which would allow us to make the effects of musical listening in small groups as measurable as possible and to validate the previously stated observations on which this method is based.

The subjects
The total number of non self-sufficient residents in the two structures when the research activity started was 667.
From the total population 90 people were selected, divided into 18 "small groups" (one for each floor) of five listeners each. Due to some leaving shortly after the activity started, the number of subjects dropped to 85.
The selection of these subjects, who we will define as "regulars", was based on three criteria:
1) the subjects’ desire to participate in group musical listening observed in a trial period in which all of the residents were invited to join informal listening groups;
2) the recommendation of reference figures in the home (doctors, nurses, health workers, activity staff and physiotherapists);
3) the psychologist’s opinion regarding the relevance of the project for each subject: those who were able to stay in the group without causing disturbance and who could communicate their feelings were included.
Each of the 18 groups was made up of two subjects with good cognitive abilities, two with medium abilities and one seriously debilitated. In this way the percentage trend of the entire population was broadly imitated in each group. To define the three categories it was established that the subjects who had cognitively correct language consistent with the context in which they were expressing themselves had "good cognitive ability"; those who showed some deficiency in this area fell into the "medium cognitive ability" category; while those who demonstrated a confusional state were "seriously debilitated".
Right from the first meetings other residents who were not part of the selected sample took to coming to the rooms where the listening music therapy was being carried out of their own accord. Unlike the "regulars", however, they were not involved in the period of discussion which followed listening to pieces of music and thus they did not actively participate in the activity.
Taking into consideration the most regular, a group of 77 residents was created who were defined as "spectators".
Finally, a further sub-sample of 55 residents from the 85 "regulars" was chosen. Those chosen had the clearest diagnoses in relation to four pathological categories: depressives, mild dementia patients, intermediate dementia patients, advanced dementia patients, to which a control group of so-called normal residents was added. Thus there were 11 subjects in each group who respectively made up the five "pathological groups" whose details have a l ready been separately dealt with.

Procedure and material
The activity took place on a weekly basis for each g roup, lasted an hour and saw the five subjects positioned around a hi-fi system to listen to six musical selections (the same for all 18 groups), each one followed by a general discussion. The activity took place in the small lounge that on each floor.
The "spectators" positioned themselves behind the "regulars" who formed a front row in a semicircle around the psychologist. He used pieces of music chosen according to the pleasure and knowledge that the clients had shown during the six months p revious to the experience itself. The CDs used contained orchestral, opera and folk music.

Survey and measurement method
To tackle the problem of verifying the music therapy activity we thought it appropriate to carry out a screening in four directions:
• study of the characteristics of the sample
• the characteristics of the pieces of music
• the behavioural reactions of the subjects
• the relationships between these and their neurological pathology.
We tried to collect data quantitatively as much as possible in the hope of identifying where the music acted and what changes it produced. The aim was to obtain information in order to be able to construct more specific and more reliable measurement tools in future.

The survey form
Creating a tool that could measure the effect of listening music therapy meant pinpointing certain behavioural signs which were considered crucial, verifying the possibility of ascertaining their changes over time and, finally, managing to create a scale which explained these variations.  
Within this framework, four indices were chosen to be treated as independent variables, plus a fifth with more complex connotations. The variables are the following:  
1. attention (this depended on the ability to concentrate on musical, verbal and visual stimuli during the activity);
2. socialisation (this depended on the way, voluntary or induced, in which the subject relates to the others as regards the theme under discussion;  
3. verbal behaviour (this measured the relevance of speech in relation to the context);
4. singing behaviour (this measured the more or less active participation in the singing). As a fifth variable that particular state of mind that is mood was measured as an expression of the "perception of well-being" (which depended on physical and psychological factors, linked to each other recursively). Thus it was not out of the question that, by modifying the psychological, the disturbance of the equilibrium would also involve the physical. The "mood" variable was measured by consulting the subjects themselves about it and comparing their sensations with the observations of the experimenter. In forming his opinion, the experimenter took into consideration the degrees with which the indexes varied (from vivacity in socialisation to emotional passion in the singing, also interpreting facial expressions when possible).
Each index was evaluated with three levels of assessment. For attention these were dozing, inattentive, attentive; for socialisation: nothing, if approached, voluntary; for sound behaviour: doesn’t sing, hints at singing, sings; for verbal behaviour: doesn’t speak, speaks incoherentl y, speaks; for mood: bad, normal, good. Shortly after the start of the first piece of music the subjects were attributed a position in the three - grade scale for each of the five variables. Then, during the meeting, there could be variations (positively or negatively). Table 1 shows the form that was used. If we consider the first subject, for example, we can see that there were no changes in the pretest and post-test phases for the attention variable, whereas positive changes were shown in the other variables.

The music
In the same form we thought it appropriate to register the level of knowledge of the title, of the melody and the agreeableness of the six pieces proposed each week to the whole sample of subjects, following a set theme. The theme of the six pieces was related to the progression of the holidays and special occasions of the calendar year and aimed to stir up memories, emotions, places, events and people related to these. For example, over Christmas we listened to seasonal music, the week after the New Year’s waltzes, and so on. The discussion that followed the listening mainly concerned the theme recalled by the music, and the residents were stimulated by precise questions which were then re c o rded in a kind of "semantic grid".
In order to define the sample of pieces of music, during the six months preceding the actual observation the 85 subjects were asked to remember as many titles as they could. The result was modest given that the number of pieces remembered was 36 (52% of subjects hadn’t remembered any). Furthermore, only three subjects had remembered at least two titles and two subjects had got to four.
The pieces were then divided into "Popular Songs", "Opera Arias", "Opera Overtures" and "Symphonies", and the percentages shown in fig. 1 resulted from this.

This percentage was then maintained in choosing a musical sample more extensive than 86 pieces. Ten of these were repeated to check the stability of the measurement: the result was positive because the pieces repeated after some time were assessed very uniformly by the 85 subjects, who thus showed themselves to be sufficiently reliable judges. Table 2 shows how the six pieces of music were "registered" during the meetings.

Greater pleasure was shown by the 85 listeners for the Popular Songs, followed by the Opera Arias, the Opera Overtures and finally the Symphonies. This trend also continued for knowing the melody and the title.
The percentage of subjects who remembered the title was lower than those who knew the melody and this latter was the result which correlated most with the agree ableness.
Spontaneously remembering the titles of pieces was very limited compared to when they were elicited by the experimenter. In total the entire sample of subjects spontaneously remembered 36 titles; more than half of these (67%) were then proposed during the music therapy encounters. Most of these are Popular Songs and the rest are Opera Arias. More than 40% of the subjects did not remember any music titles of their own accord. The ability to spontaneously access their musical memory in its typically cognitive perspective (remembering titles, singers, musicians) showed itself to be much reduced, in contrast to what happened during the music therapy activity given that a good part of the 86 pieces proposed were appreciated and remembered both in title and in melody by a high percentage of people. This suggests that listening to a melody also recalls its cognitive content and this all influences the agree ableness and, consequently, mood if it is true that a pleasant experience acts agreeably on our perception of well-being.

RESULTS
The 667 subjects who re p resented the total population on 01-06-1996 were divided, as already described, into three groups: the "non-attendees" (505), the "spectators" (77) and the "regulars" (85).
After 18 months (in December 1997) the percentage of deaths was checked and somewhat evident differences between the three groups were discovered. The group of "non-attendees" had a mortality rate of 31.48%, the group of "spectators" of 27.27% and the group of "regulars" of 18.82% (see fig. 2).

Figure 2. The graph shows the mortality rate in the three groups and the relative percentages (regulars 18.82%, spectators 27.27%, non-attendees 31.48%) between June ’96 and December ‘97

Given the criteria with which the subjects were chosen, the composition of the group of "regulars" was representative of the total population as regards cognitive abilities; it is thus legitimate to ask ourselves the nature of such evident percentage differences. One hypothesis is that the effects of the listening music therapy had an influence of the level of physical health of the people by virtue of the psychological processes. However, the psychological "data" available to us were not strict enough to allow us to rashly infer connections, but we can present the results from the elaboration of the measurement forms described above and from the survey about the music and the relative memories.

The behavioural variables 
It now remains to be understood whether mood really improved during the music therapy. To do so we can interpret the elaborations of the data from the measurement form related to the five behavioural variables.
The graph in fig. 3 shows the average trend of the sample of 85 subjects during the sixteen weeks of observation. The difference between each pair of points re p resents the effect of the musical activity. 

Figure 3. the two lines show the average variations of the behavioural indices of the entire sample from the beginning (bottom line) to the end (top line) of each encounter 

The trend of the two curves over the course of the sixteen weeks of observation is not linear. 
Their considerable difference compared to the average score highlights, instead, how the entire sample of subjects clearly improves from the beginning to the end of each individual encounter. This testifies an clear effect during the activity, though it does not allow us to ascertain long-term improvements.
Moreover, studying the separate averages of the five variables, it was observed that the trends of two of these (verbal behaviour and attention) are very similar, whereas the trends of socialisation, singing behaviour and mood, also very similar to each other, better represent the trend of the overall average of the five variables. These two factors were called "cognitive" and "emotional", presuming that mood is more closely connected to the latter, though also useful for stimulating the former in the form of motivation to be in a group, to socialise and thus to remember the contents of the theme under discussion (music, places, people, etc.).

Figure 4. The thicker line "a" represents mood, "b" the emotional variable, "c" the cognitive variable, over the 16 weeks of observation 

The above is illustrated in the graph in fig. 4 where the curve with line "a" represents the aver-age of the "mood" variable, line "b" the "emo-tional" variable (socialisation and singing behav-iour) while line "c" the "cognitive" variable (attention and verbal behaviour). Each of the three curves shows the variation in the score over the sixteen weeks (final score minus initial score).

The pathological groups 
The elaboration of the data for the four groups of 11 subjects classified by doctors according to their pathologies (depression, mild dementia, intermediate dementia, advanced dementia) highlighted some interesting aspects: during the activity the mood of the depressives improved more and the attention of the dementia patients remained high. 
A further elaboration with statistical methods based on autorefractive neural networks (Pandin and Didona’, 1996; Angi, Sturaro, Pandin and Didone’, 1997) showed that the behaviour during the activity (measured with the form) could differentiate membership of pathological groups as it was defined by the doctors. The most discrimination is found between the "advanced dementia patients" and the "normal" residents, whereas the attribution of pathological category for the others becomes less certain in relation to doctors’ opinion. Furthermore, mood and attention appeared to be connected, which could indicate that interest in music improves one’s well-being.
The difficulty in correctly attributing the depressed subjects and those with mild forms of dementia to their respective categories suggested that some aspects of these pathologies are inter-woven, making the interpretation more complex: the advance of dementia accompanies depressive states, but maybe the atmosphere that one lives in a retirement home does not give the resident the motivation to keep his cognitive potential efficient.

Conclusions
The data that emerged from the investigation lead us to believe that the music had beneficial effects on the listeners. The benefits are related to the perception of well-being expressed by mood. It seems that the mnemonic, emotional and cognitive activations that music stirs played a fundamental role.
These effects were revealed over the course of the activity; it will be interesting to verify their duration in time, even if the elaborations of the data on the behavioural variables and, above all, on the percentage of deaths of the listeners would seem to testify this. As already stated, the difference between the three groups of subjects showed itself to be surprisingly high and so the results of this pilot experience are worth further development and verification. By saying this we do not intend to claim that music therapy is the only factor responsible for the lower mortality rate, but to acknowledge that an emotional variation capable of positively stimulating the perception of well-being of an individual can also affect maintenance of their cognitive abilities and their general physical condition.
In this way, we can state that music prolongs life by stealing the beloved areas of the memory from the darkness of oblivion. 

Numerous studies on musical perception, memory and emotion, dimostrated that the uman brain elaborates music in a modular way

Musical memories are laden with emotional content which is very easely retrieved in memories even in those who are seriously debilitated from a cerebral point of view

Angi M., Sturaro M., Pandin M., Didone G.
Screening for Amblyopia with a hand-held autore fractor in pre-school children : automated data interpretation using a neural network system, Perception, 26, 1997, pg. 770 - 771

Ciompi L.
L’ipotesi della logica affettiva (The hypothesis of affective logic), Le Scienza quaderni (The Sciences journals), n^o. 82, February 1995, pg. 58-68.

Goleman D.
Intelligenza Emotiva (Emotional Intelligence), Milan, Rizzoli, 1997 (or. ed. 1995).

Grof S.
Realms of Human Conscious, London, Souvenir Press, 1975.

Le Doux J. E.
Emozioni, memoria e cervello (Emotions, memory and brain), Le Scienze (The Sciences), 1994, no. 312, August, pg. 32-39.

Le Doux J. E.
Il cervello emotivo (The emotional brain), Milan, Baldini & Castoldi, 1998 (or. ed. 1996).

Leaf A.
Invecchiamento e longevita’ (Ageing and longevity), Le Scienze (The Sciences), no. 73, September 1974, pg. 32-41.

Mancia M.,
Sull’esperienza musicale: riflessioni neurofisiologiche e psicoanalitiche (On musical experience, neuro physiological and psychoanalytical considerations), in C. Baconcini, G. Manarolo, R. Rossi, F. Scarsi (edited by) Musicoterapia e turbe psichiche (Music therapy and mental disorders), Genoa 1989, pg. 14-22.

Pandin M., Didone G.
Chaos in information processing: Stimulation of a biological learning process by time evolution in an unsupervised neural network, International Journal of Biforcation and Chaos, Vol. 6, No. 1, 1996, pg. 203-210.

Scapagnini U. – Scapagnini G.
La manutenzione della vita (The maintenance of life), Milan, Mondadori, 1997.