-
Development of the University of
Pennsylvania Smell Identification
Test:
A Standardized Microencapsulated Test of Olfactory
Function
宾夕法尼亚大学气味识别能力测试
(
UPSIT
)的开发:
嗅觉功能的标
准化微胶囊测试
PHYSIOL BEHAV
32(3) 489-502, 1984
The
development of the first standardized
Over
1600
subjects
participated
in
five
experiments.
In
Experiment
1,
50
microencapsulated
odorants
were
rated
as
to
their
intensity,
pleasantness,
irritation,
coolness, and
familiarity, and two procedures for releasing them
were compared. In
Experiment
2,
the
results
of
the
first
experiment
and
other
data
were
used
in
the
development of the test,
which was administered to a large number of
subjects. Using
multiple regression
analysis, scores on this test were shown to be
significantly related
to the subjects'
gender, ethnic background, and smoking behavior.
Average test scores
decreased as a
function of age, with the greatest decline
occurring between the sixth
and tenth
decades of life. These age-related changes were
not correlated with scores
on
the
Wechsler
Memory
Scale.
Women
performed
better
than
men
within
all
age
categories. In Experiment 3, the test
was shown to differentiate between subjects with
known
olfactory
disorders
(e.g.,
KaUmann's
syndrome;
Korsakoff's
syndrome)
and
normal
controls,
and
to
reliably
detect
persons
instructed
to
feign
total
anosmia.
In
Experiment
4,
the
test-retest
reliability
was
established(6-month
interval;
r=0.918,
p<0.001), and in Experiment 5 the test
was shown to correlate thresholds with odor
detection (r=-0.794, p<0.001). This
self-administratered test now makes it possible to
rapidly
and
accurately
assess
general
olfactory
function
in
the
laboratory,
clinic,
or
through the mail without
complex equipment or space-consuming stores of
chemicals.
本文描述了首个标准化
“刮
—
嗅”
式嗅觉测试方法
的建立。
在开展的五个实验中,
招募的受试者累计超过
1600
名。在实验
1
中,对
50
种微囊气体就其强烈度、愉悦
度、
刺激度、
冷暖度和熟悉度五种类型的感知进行评分,
p>
并对两种释放气味的方
式进行了比较。
在实
验
2
中,
基于实验
1
的结果及其他数据建立检测方法,
并将其
施用于大量的受试者。采用多元回归分析,结果表明测试得分同受试者的性别、
种族背景、
以及吸烟行为存在显著的关联性。
测试平均得分随年
龄的衰老而降低,
在
60
至
70
岁之间降幅达到最大。
这些与年龄相关的变化
与韦氏记忆量表评分不存
在相关联性。在所有的年龄段中,女性的表现均优于男性。在实
验
3
中,测试结
果表明这种检测能有效
的区分已知嗅觉障碍(例如,
KaUmann
综合征;柯萨可
夫氏综合征)
和正常对照组的受试者,
并能可靠地识别出伪装嗅觉完全缺失的人
员。在实验
4
中,建立了重测信度(
6
个月时间间隔;
r =
0.918
,
p <0.001
)。
在
实验
5
中,检测结果表明了阈值与气
味识别的关联性(
r =
-0.794
,
p <0.001
)。
现
在这种自我测试的方式使得快速而准确地评估一般嗅觉功能成为可能,
无论是在
实验室、
还是诊所,
甚至是通过邮寄的方式,
无需复杂的设备和占用空间来存储
的化
学品。
THE
sense
of
smell
largely
determines
the
flavor
of
the
foods
we
ingest
and
the
beverages we savor, and serves as an
important early warning system for the detection
of
fire,dangerous
fumes,
leaking
gas,
spoiled
foods,
and
polluted
environments.
Despite
these
important
functions,
few
physiology
or
medical
textbooks
discuss
procedures for
evaluating this sense, and most fail to emphasize
the fact that olfactory
disorders
commonly
occur
as
a
result
of
accidents,
disease
states,
medical
interventions, aging, and exposure to a
number of environmental pollutants [14, 20,
29,
59].
Furthermore,
it
is
rarely
mentioned
that
smell
disorders
can
serve
as
important diagnostic signs of a number
of serious diseases and anomalies, including
ones
related
to
the
ontogeny
of
the
hypothalamus
and
pituitary
[10,
16,
35,
50],
intracranial neoplasms
[3,22], and temporal lobe epilepsy [5, 7, 46].
嗅觉在很大程度上决定了我们所摄取食物和品尝饮料的味道,并作为探测火灾、
危险气体、
气体泄漏、
食品腐坏及环境污染的一个重要
的早期预警系统。
尽管有
着这些重要的功能,
< br>然而很少生理学和医学教科书上有讨论到评估这种感知能力
的方法,更为重要的是
忽略了一个事实,那就是嗅觉障碍常常是由于意外事故、
疾病、医疗干预、衰老和暴露于
多种环境污染物而导致的
[14
,
20
,
29
,
59
]
。此
外还很少提到,嗅觉障碍其实是可以作为许多严重疾病和
异常的重要诊断标志,
包括那些涉及到下丘脑和垂体
[10
p>
,
16
,
35
p>
,
50]
、
颅内肿
瘤
[3
,
22]
和颞叶癫痫
[5
,
7
,
46]
等相关的疾病。
Such omissions do not stem
from a want of persons evidencing olfactory
problems.
For example, data from the
National Ambulatory Medical Care Survey suggest
that,
for 1975 and 1976 combined,
435,000 visits to physicians' offices occurred in
which a
major presenting complaint was
chemosensory in nature [47]. Leigh 137] noted a
7.2%
anosmia
rate
in
1000
consecutive
head
injury
cases
admitted
to
a
British
military
hospital during WWII and Sumner [53]
reported a 7.5% incidence of post-traumatic
anosmia in a study of 1167 head injury
cases observed at the Leeds General Infirmary.
Approximately
20%
of
the
tumors
of
the
temporal
lobe
or
lesions
of
the
uncinate
convolution produce some form of
olfactory disturbance [25], and it is now believed
that about a third of schizophrenic
patients exhibit olfactory hallucinations of one
type
or another [8, 45, 49].
这样的疏漏并非出于嗅觉研究人员的本意。
例如,
来自全美国
门诊医疗护理调查
的数据显示,
1975
年和
1976
年,
在
435000
名门诊患者中,
一个主要的现病史是对<
/p>
自然界中化学物的感受
[47]
。
Leigh [37]
指出在世界二次大战期间,
被送往一家英
国军事医院的
1000
名持续脑部损伤病例中有
7.2
%的嗅觉丧失率。
Sumner
[53]
则
在对利兹综合医院收治的
1167
名颅脑损伤病例的研究中,
发现存在有
7.5%
的创伤
后失嗅率。约
20
%的颞叶肿瘤或海马回钩病变会
产生某种形式的嗅觉紊乱
[25]
,
现
在普遍认为,
有大约三分之一的精神分裂症患者表现出某种嗅幻觉
[8
,
45
,
49]
。
A major reason for the dearth of
interest on the part of human-oriented basic
scientists
in
this
important
sensory
system
has
been
the
lack
of
a
practical,
objective,
and
standardized means for testing its
function. Unlike the testing of the visual,
auditory,
and
vestibular
systems
(where
standardized
tests
are
widely
applied),
the
testing
of
olfaction
varies
considerably
from
laboratory
to
laboratory.
Clinically,
such
tests
range
from
simply
asking
a
patient
whether
a
smell
problem
exists
to
the
administration of
lengthy and impractical threshold tests for which
sound normative
data are typically
lacking.
一向以人为本的基础研究科学家,
对于这一
重要感官系统却缺乏研究兴趣的一个
主要原因就是缺乏一种具备可操作性、
客观性和标准化的嗅觉功能评估方法。
不
像视觉、听
觉和前庭系统(在这些领域都有被广泛应用的标准化测试手段),嗅
觉功能的检测在不同
实验室之间存在着很大的差异。
而在临床上,
检测手段则从
p>
简单的问询患者是否存在嗅觉问题到漫长而不切实际的阈值检测,
长
期以来都缺
乏足够充分的规范化数据。
The present paper describes
five experiments leading to the development of the
first
standardized
'n
sniff
test
of
olfactory
function--the
University
of
Pennsylvania Smell Identification Test
(UPSIT
and
2,
the
empirical
bases
upon
which
the
stimuli
and
response
alternatives
were
selected for inclusion
in this test are described, along with studies of
the influences of
variables such as the
age and gender of the subject on the test scores.
In Experiment 3
an evaluation is made
of the utility of the UPSIT in discriminating
between persons
with
and
without
olfactory
dysfunction,
as
well
as
persons
instructed
to
feign
total
anosmia.
In
Experiment
4 a study of
the UPSIT's
test-retest
reliability
is
presented,
and
in
Experiment
5
a
correlation
determined
between
the
UPSIT
test
scores
and
measures from a traditional odor
detection task.
本论文描述了五个实验,
进
而建立了第一个标准化的
“刮
—
嗅”<
/p>
式嗅觉功能测试
方法
——
宾夕法尼亚大学气味辨别试验(
UPSIT
,见图
1
< br>)。在实验
1
和
2
中,
对本测试中纳入的刺激源以及设定的应答选项所依据的经验基础进行了说
明,
同
时研究了相关变量(如受试者的年龄和性别)对测试得分
的影响,在实验
3
中,
对采用
UPSIT
区分嗅觉正常者和嗅觉障碍者的效果进行了评估,
包括识别事前故
意安排的伪装嗅觉完全丧失者。在实验
4
中,提出了
UPSIT
的重测信
度,在实验
5
中,探明了
UPSIT<
/p>
测试得分同传统的气味鉴别手段之间的关联性。
FIG. I. Picture of the
University of Pennsylvania Smell Identification
Test. Note
the response columns on the
last page of the booklets.
EXPERIMENT I
实验
1
Experiment
1
had
four
main
goals.
The
first
was
to
quantitatively
establish,
in
subjects
with
no
olfactory
dysfunction,
the
perceived
intensity,
pleasantness,
familiarity,
coolness-
warmth,
and
irritation
of
50
Microfragrance
TM
samples
of
potential
use
in
a
standardized
olfactory
test.
(Microfragrance
TM
is
a
registered
trademark
of
the
3M
Company,
Minneapolis,
MN.)
Such
data
provided
basic
information as to the suitability of
microencapsulated odorants for human testing, as
well as a basis for eliminating stimuli
with clear problems of identifiability,
irritation,
or
intensity
from
the
final
stimulus
set.
The
second
goal
was
to
determine
whether
such
ratings
were
differentially
influenced
by
two
simple
means
of
releasing
the
stimuli (scratching the odorized
surface with a pencil tip or with No. 120
sandpaper)
and,
if
so,
whether
one
procedure
was
preferable
to
the
other.
A
third
goal
was
to
establish if a sex difference occurs in
the obtained intensity and pleasantness ratings
of the stimuli, as expected from
studies which suggest that women rate a number of
odors as more intense and less pleasant
than do men [13, 17-19]. The fourth goal of
this
study
was
to
ascertain
the
relative
identifiability
of
the
Microfragrance
TM
samples
when
no
verbal
or
written
cues
were
provided
as
to
their
identity.
This
information, in
conjunction with the familiarity ratings obtained
in this study and data
collected in the
next experiment (Experiment 2), was subsequently
used to eliminate
stimuli that were
difficult to identify.
实验
1
p>
有四个主要目标。第一,针对没有嗅觉障碍的受试者,对应用于标准化嗅
觉测试的
50
种
Microfra
grance
TM
候选气味样本,
就其强烈度、愉悦度、熟悉度、
冷
—
暖
度和刺激度五种类型的感知情况,建立定量评估方法(
Microfragrance
TM
为
3M
公
司注册的商标)。这些数据将为用于人体试验的微胶囊气味的适用性提
供基本的信息,<
/p>
同时也为剔除那些存在明显可识别度、
刺激度或强烈度问题的刺<
/p>
激物进而获得最终的最佳刺激源组合提供基础。
第二,
评估这样的得分是否会因
刺激物两种简单的释放方式(用铅笔尖或
120
号砂纸刮擦气味剂的表面)而产生
差异化
的影响,如有影响,是否其中一种方式更优于另一种方式。第三,对于所
获得的刺激源的
强烈度和愉悦度的评分,
判定是否存在性别差异。
因为之前已有
研究表明,
相比于男性,
女性更倾向于
将更多的气味归为强烈,
而归为愉悦的很
少
[13
,
17-19]
。第四,确
定
Microfragrance
TM
样品在没有口头或书面身份信息时
所具有的相对可辨识性。
这
些信息,
同在本研究中获得的熟悉度评分以及在下一
个实验(实
验
2
)中收集到的数据,将随后被一起用来剔除那些难以识别的
刺激
源。
Subjects
试验对象
Twenty-six men and 26 women (mean
age=24.87 years, SD=5.52 years, range=18 to
40 years) of excellent health and with
no apparent olfactory problems were selected
from a subject pool maintained by the
Clinical Smell and Taste Research Center. All
but
10 were non-smokers and,
of these10, only two
smoked more than 1
pack/day.
Forty-four were White
Americans and eight were Black Americans, and the
majority
were
college
students
at
the
University
of
Pennsylvania.
None
had
previously
participated in chemosensory studies.
Each subject received $$10.00 for participation.
p>
男女受试者各
26
名(平均年龄
=24.87
岁,标准差
=5.52
岁,年龄范围
=18-40
岁),
< br>身体健康,
没有明显的嗅觉问题,
所有受试者均从临床嗅
觉味觉研究中心的受试
者信息库中选取。
所有受试者中,
有
10
人为吸烟者,
其中只有两人每天吸烟量超
过
1
包。<
/p>
44
人为美国白人,
8
< br>人为美国黑人,
大部分受试者是为宾夕法尼亚大学在
校学
生。没有任何人以前曾参加过有关化学感受的研究。每位受试者均发放
$$10.00
p>
的试验补贴。
Odorants
气味
The 50 stimuli
used in this
and the subsequent
experiments were
chosen from
over
one
hundred
Microfragrance
TM
samples
available
from
the
3M
Company
(Minneapolis,
MN)
on
the
basis
of
several
criteria:first,
that
their
odors,
as
judged
from
preliminary
samples
sent
to
the
experimenters,
were
a
reasonable
first-order
approximation of
the descriptive label given to them by the
manufacturer; second, that
they
spanned
a
number
of
previously-established
qualitative
odor
classes
[1,28],
including
disparate
sectors
of
a
multidimensional
perceptual
space
developed
using
microencapsulated
odorants [9]; third, that they included stimuli
composed of single
as well as multiple
components (e.g., licorice is typically a single-
component odorant
made up of anethole,
whereas chocolate is a multiple-component odorant
made up of
a
number
of
chemicals),
given
the
possibility
that
the
olfactory
system
codes
information
on
the
basis
of
a
multiple
profile-multiple
receptor
site
process
[43];
fourth,
that
the
majority
evidenced
no
or
minimal
ability
to
stimulate
non-
CN
I
intranasal
or
pharyngeal
chemosensory
systems
(e.g.,
CN
V);
and
fifth,
that
a
few
trigeminal
stimulants
(e.g.,menthol)
be
included
to
allow
detection
of
at
least
some
types of malingerers
[16]. The odorants were embedded in 10--50 /zm
plastic capsules
coated onto adhesive
backed labels, as described elsewhere [41].
用
于
本
次
< br>实
验
及
随
后
实
验
的
5
0
种
刺
激
源<
/p>
均
来
自
3M
p>
公
司
,
是
从超
过
100
种
Microfragrance
TM
样品中根据以下标准选取而来的:
第一,
所使用的气味是从发送给实验人员的样品中初步筛选出来
的,
是对生产商
所提供的描述性标签的一种合理的一阶近似;<
/p>
第二,
选取的气味覆盖了大规模之
前已定
性确立的气味类别
[1
,
28]
,包括利用微胶囊气味剂开发的多维感知空间
的不同部分
[9]
;
第三,
刺激源可以
是单一成分,
也可以是多种成分组成
(例如,
< br>甘草是典型的由茴香脑构成的单一成分气味剂,
而巧克力是由多种化学成分构成<
/p>
的多组分气味剂)
,
这主要是考虑到嗅觉
系统感知信息是基于多重信号—多重受
体位点这样一种信号传递通路
[43]
;第四,绝大部分气体已被证明不能或最小
p>
程度的刺激非
CN I
鼻内或咽化学感受系统(例如,
CN V
);第五,需纳入若干
种三叉神经刺激物(例如,薄荷醇)以便能检测出至少某些类型
的诈病
[16]
。所
有气味物质均包埋
在
10
–
50um
大小的塑料胶囊中,
贴附在背面具有粘性的标签纸
上,如文
献
[41]
所述。
Procedures
流程
The 52
subjects were divided into 13 groups of four
apiece. Within each group there
were
two males and two females of similar age. The
average age was approximately
21 for
six of the groups, 25 for three of the groups, 28
for two of the groups, and 31
and 38
for the other two groups. The eight black subjects
formed two of the groups.
Within each
group, one member of each sex released the
microencapsulated odorants
by
scratching the labels with a 2 by 1 inch strip of
No. 120 sandpaper, whereas the
other
used the tip of a No. 2 lead pencil. The
instructions were read to each subject by
an experimenter who monitored the test
session to insure that the correct procedures
were followed.
5
2
名受试者被分成
13
组,每组
4
人。每组有年龄相仿的两名男性和两名女性。其
中有
6
组的平均年龄约为
21
p>
岁,
3
组的平均年龄约为
< br>25
岁,
2
组的平均年龄约为<
/p>
28
岁,另两组的平均年龄分别约为
31
岁和
38
岁。
8
个黑人组成了其中的两组。每组
中,
男女各一名成员负责通过使用
2*1
英寸大小的细条状的
120
号砂纸刮擦标签纸
来释放胶囊中的气味
,而另两名男女成员则使用
2
号铅笔尖刮擦。由负责监督试
p>
验过程的实验人员负责向受试者宣读相关试验说明,
确保试验按照预
定的程序进
行。
Each subject rated, in individual hour-
long test sessions, all of the 50 Microfragrance
TM
samples on five 9-point
category
scales with
the
following
adjectives
serving
as
anchors at their
extremes: very weak---very strong; very
unpleasant
—
very pleasant;
non-irritating--very
irritating;
very
unfamiliar--very
familiar;
and
very
cool--
very
warm. One 3 by 1/2 inch
Microfragrance
TM
sample
label was located at the bottom of
each
page of an 81/2 by 11 inch test booklet which
contained each of the five 6-inch
long
scales. The 50 pages were completed in the order
in which they were presented.
To
control for position response biases of the
subjects, as well as for order effects in
the presentation of the scales
themselves, both the order in which the scales
appeared
on
a
page
and
the
position
of
the
adjectives
on
a
given
scale
(e.g.,
very
unpleasant
—
very
pleasant vs. very pleasant--very unpleasant) were
counterbalanced
across
subjects,
as
in
earlier
work
[16].
In
addition,
the
order
in
which
the
stimuli
were
presented was also counterbalanced across
subjects. The subjects were allowed
to
work
at
their
own
rate,
but
were
required
to
take
a
10
minute
break
halfway
through
the
task
(i.e.,
after
the
25th
MicrofragranceTM
sample).
A
label
could
be
repeatedly
scratched,
as
needed,
before
moving
to
the
next
odorant,
although
returning to
previous odors was not allowed. Following
completion of the scales on a
given
page,
the
subjects
were
asked
to
write
down,
to
the
best
of
their
ability,
the
identity of the odor presented on that
page.
在长达一个小时的测试过程中,每名受试者需对所有
50
种
Microfragrance
TM
样品
在五个
9
< br>分式范畴量表上进行打分,表上每个形容词的极限作为两端锚点:
非常弱
—
非常强;非常不愉悦
—
非常愉悦;无刺激性
—
非常刺激;
很陌生
—
非
常熟悉;
非常冷
—
很温暖。
一个
3*
1/2
英寸大小的
Microfragrance
TM
样品标签置于
8.5*11
< br>英寸大小的测试手册的每一页的底部,每页测试手册均含有五个
6
英寸长
范畴量表。
50
页按照
原有的顺序依次完成。
为了控制受试者对于位置的反应偏差
以及
范畴量表本身的顺序可能产生的影响,
五个范畴量表出现在一张纸上的顺序
和每个范畴量表上形容词的位置
(例如,
非常不愉悦
—
非常愉悦
与
非常愉悦
—
非常不愉悦)在受试者中进行交叉平衡分布,这在早期的研究中已有使用
[16]
。
此外,
对于刺激
源的施加顺序也采取在受试者中进行交叉平衡的方式。
受试者被
允许可以按照他们自己的速度进行测试,
但是要求在整个测试任务过程中中间必
须休息
10
分钟(例如,完成了
25
种
Microfragrance
TM
样品后)。每种标签可以根
据需要进行重复刮擦
,
直至转移到下一个气味剂,
此时不允许再返回到前一个气
p>
味剂。
在完成了每一页的范畴量表后,
受试
者将会被要求尽他们最大的能力写下
该页所附气味剂的身份。
Data Analysis
数据分析
To determine
if the procedure used
for
releasing
the stimuli
or if the
gender of the
subjects influenced the psychological
ratings, the scale values of each attribute were
analyzed by a 2 (sandpaper vs. pencil)
by 2 (male vs. female) by 50 (Microfragrance
TM sample) analysis of variance
(ANOVA). A separate ANOVA was performed for
each of the five attributes since it
was not clear that the same metric was being used
for all scales (e.g., warm-cool and
unpleasant-pleasant are bipolar scales, whereas
the
others are unipolar, although not
necessarily of comparable
scale).
为了判别所使用的刺激源释放方式和受试者
的性别是否影响到了心理评分,
每一
参数的范畴量表值采用
p>
2
(砂纸与铅笔)
*2
(男性与女性)
*50
(
Micr
ofragrance
TM
样品)的方差分析模式
(ANOVA)
进行分析。五个参数的每一参数都进行一次独
立的
ANOVA
分析,
因为
并不能明确相同的度量衡适用于所有的范畴量表
(例如,
暖
p>
—
冷
和
不愉悦
—
愉悦,
是双极范畴量表,
而其他的则是
单级的,
并不一定具
有可比性)。
To
quantify
the
identifiability
of
the
stimuli,
as
indicated
by
the
written
responses
provided by each
subject, a point was given for every response that
was identical to
that
of
the
manufacturer's
suggested
label
(e.g.,
licorice
for
licorice).
In
addition,
a
point
was
given
for
responses
indicative
of
products
containing
the
odorant
as
the
major
component
(e.g.,anise
for
licorice).
A
half
point
was
given
to
less
specific
responses which, nonetheless, were at
least remotely related to the odor (e.g., candy
for
licorice).
The
sum
of
the
points
across
the
52
subjects
served
as
an
odorant's
index.
In
the
present
context,
this
simple
scoring
system
seemed
preferable
to
more
complex
ones
noted
in
the
literature(e.g.,
[ll]),
in
that
little
subjectivity was required in assigning
responses to the specific categories.
为
了对刺激源的可识别度进行量化,
对于每个受试者所给出的书面应答,
< br>每一应
答如果和生产商建议的标签是一致的,则给予
1<
/p>
分(例如,甘草
对应
甘草)。
此外,如果应答指明的是该产品所包含的气味剂主要成分,也给予
1
< br>分(例如,
茴香
对应
甘草)
。
对于特异性不强的应答,
但仍与目标气味剂具有一定关联性<
/p>
的,则给予半分(例如,糖果
对应
甘草)。每种气味剂的
52
名受试者总得分作
为该气味剂的
―
可识别性指数
‖
。
p>
就目前来看,
这种简单的评分系统似乎要优于其
他文献报道的更为复杂的系统(例如,
[11]
),在该文
献中,在对应答进行特异
性归类时具有一定的主观性。
RESULTS
结果
The overall
mean ratings for each of the 50 MicrofragranceTM
samples are presented
in Fig. 2 for all
five psychological attributes. It is evident from
this figure that none of
the stimuli
was rated so extreme on any of the continua as to
warrant its immediate
exclusion from
consideration in the final stimulus set.
Furthermore, it is also apparent
from
this figure that the stimuli generally differed
among one another on each of the
perceived
attributes,
as
indicated
by
significant
Microfragrance
TM
sample
main
effects for each of the five analyses
of variance (all ps<0.001; Strength F(49,49)=7.60;
Familiarity F(49,49)= 15.12; Irritation
F(49,49) =28.09; Pleasantness/Unpleasantness
F(49,49)=37.68; Warm/Cool
F(49,49)=9.81).
50
种
Mi
crofragrance
TM
样品中,每一种样品的五个参数
的总体平均得分情况如图
2
所示。
从该
图可以很明显的看出,
没有任何刺激源在任何参数上得分非常极端,
以至于可以将其直接排除,
不纳入最终刺激源组合的考虑范围。
此外,
从该图也
很
明
显
的
看
出
< br>,
每
个
感
知
参
数
在
各
刺
激
源
间
p>
普
遍
都
各
不
相
同
,
这
也
从
Microfrag
rance
TM
样品对于五个方差分析都具有明显的主效应可
以看出(所有
ps<0.001;
强度
F
(
49,49
< br>)
=7.60;
熟悉
F
(
49,49
)
=1
5.12;
刺激
F
(
49,49
)
=28.09;
愉悦
/
不愉悦
F
(
49,49
)
=37.68;
暖
/
冷
F
p>
(
49,49
)
=
9.81
)。
FIG. 2. Mean category ratings given to
50 microencapsulated odorant strips using No.
120 sandpaper (left half of each
vertical bar) or the tip of a No. 2 lead pencil
(right
half
of
each
bar).
S=Strength
(intensity);
F=Familiarity:
I=Irritation;
P/U=Pleasantness/Unpleasantness;
W/C=Warm/Cool.
For
the
P/U
and
W/C
scales,
the
horizontal line signifies the neutral reference
point. Note, for example, the marked
unpleasantness ratings and high
irritation ratings given to onion, but not to
rose, and
the coolness attributed to
menthol.
As
expected
from
previous
work,
women
rated
the
odors,on
the
average,
as
more
intense
than
did
men
(Table
I
;
Gender
Main
Effects:
Intensity
F(1,49)=8.70,
p<0.005).
Somewhat unexpectedly, women also rated, relative
to
men, the odors as
less
cool,
less
irritating;
and
more
familiar
(Table
1
;
Gender
Main
Effects:
Cool-Warm
F(1,49)=6.80,
p=0.0t2;
Irritation
F(1,49)=
13.24,
p<0.001;
Familiarity
F(1,49)= 15.51,
p<0.001), suggesting that gender differences are
present for attributes
in addition to
intensity. Women tended to rate the odors as more
unpleasant than did
men,
although
the
results
did
not
achieve
statistical
significance,
F(!,49)=3.22,
p=0.079.
Interestingly,
the
largest
relative
difference
between
the
sexes
was
in
the
rated familiarity of the stimuli.
正如之前研究所预测的一样,
总体来说,
与男性相
比,
女性对于气味的评分更偏
向于强烈(表一
< br>;
性别主效应:强烈度
F
(<
/p>
1,49
)
=8.70
< br>,
p <0.005
)。有些出乎意
料的是,相对于男性来说,女性对于气味的评分还偏向于冷感觉小、刺激度小、
熟悉
度多(表
1
;性别主效应:冷
—
暖
F
(
1,49
p>
)
=6.80
,
p
=0.012;
刺激
F
(
1,49
)
=13.24
,
p <0.001;
熟悉
F
(
1,49
)
= 15.51
,
p <0.001
)
,
这表明,
不仅在强
烈度上,
在其他参数上性别差异也是存在的。
相比于男性,
p>
女性更倾向于把气味认为是不
愉悦的,虽然结果没有达到统计学显著
意义,
F
(
1
,
49
)
=3.22
< br>,
p =0.079
。有趣
的是
,两性之间的最大相对差异是在对刺激源的熟悉度评分上。
In
general,
the
procedure
used
to
release
the
stimuli
influenced
the
psychological
ratings.
Thus,
when
released
by
sandpaper,
the
stimuli
were
rated,
overall,
as
significantly
more
familiar
and
less
pleasant
(Table
2;
Release
Procedure
Main
Effects: Familiarity F(1,49)=7.83, p
=0.007; Pleasantness. F( 1,49) = 8.36, p <0.006).
Although not reaching the 0.05 level of
statistical significance, there was a tendency
for stimuli to be rated as less intense
and more irritating when released by sandpaper
than
when
released
by
pencil
(Table
2;
Release
Procedure
Main
Effects:
Strength
F(1,49)=3.36,
p=0.073; Irritation F(1,49)=3.86, p=0.055). No
statistically meaningful
influence of
the releasing procedure upon the cool/warm ratings
was present (Table 2;
Release Procedure
Main Effect F(I,49)= 1.13, p=0.293).
一般
来说,刺激源的释放方式也会影响心理评分。因而,当使用砂纸释放时,总
体而言,刺激
源更偏向于被评为熟悉度高和愉悦度低(表
2
;释放方式主效应
:
熟悉
F
(
1
,49
)
=7.83
,
p
=
0.007;
愉悦
F
(
1
,
p>
49
)
=8.36
,
p
<0.006
)。虽然没有
p>
达到
0.05
水平的统计学显著意义,但仍
呈现出一种倾向,那就是使用砂纸比使用
铅笔释放刺激源具有更低的强烈性和更强的刺激
性(表
2
,释放方式主效应:强
度
p>
F
(
1,49
)<
/p>
=3.36
,
p=0.073;
刺激性
F
(
1,49
)
=3.86
,
p=0.055
)。释放方式对冷
/
暖的评分并没有呈现出有统计学意义的影响(表
2;
释放方式主效应
F
< br>(
I
,
49
)
=1.13
,
p=0.293<
/p>
)。
In addition to having a direct effect
upon a number of the rated attributes, the release
procedure interacted with several of
the
other variables.
Thus,
women rated stimuli
released
by
sandpaper
as
stronger
than
those
released
by
pencil
(respective
means=6.51 and 6.26), whereas men rated
stimuli released by pencil as stronger than
those released by sandpaper (respective
means
=6.43 and 5.93) (Gender by
Release
Procedure
Interaction
F(
1,49)
=
28.51,
p
<0.001
).
Similarly,
women
rated
stimuli
released
by
sandpaper
as
more
familiar
than
those
released
by
pencil
(respective
means=6.32
and
5.92),
whereas
this
difference
was
not
apparent
for
the
men
(both
means=5.84)
(Release
Procedure
by
Gender
Interaction
F(1,49)=7.83,
p=0.007).
Some
odorants
were
rated
as
slightly
more
familiar
than
others
when
released
by
sandpaper than when released by pencil
(Fig. 2; Release Procedure by Microfragrance
TM sample Interaction F(49,49) = 1.67,
p =0.038).
除了对许多评分参数有直接影响外,
释放方式也同其他几种变量有着相互影响。
因而,
对于女性来说
,
使用砂纸比使用铅笔释放刺激源的效果更偏向于强烈
(砂
p>
纸和铅笔的平均值分别为
6.51
和
6.26
),而男性的情况则正好相反(铅笔和砂纸
< br>的平均值分别为
6.43
和
5.
93
)
(性别与释放方式间的相互影响
F
(
1,49
)
=28.51
,
p<0.001
p>
)。类似的,对于女性,使用砂纸释放的效果比铅笔偏向于熟悉度更高
(砂纸和铅笔的平均值分别为
6.32
和
5.92
),而对于男性这种差别并不明显(两
者平均值均
为
5.84
)(释放方式与性别间的相互影响
< br>F(1,49)=7.83,
p=0.007)
)。<
/p>
相比于铅笔,
当使用砂纸释放时,
一些气
味剂在熟悉度上的得分比其他一些气味
剂较高一点(
Fig.2
;释放方式与
Microfragrance
TM
样品间的相互影响
F(49,49)
=
1.67,
p =0.038
)。
Overall, there was a significant
tendency for the familiarity ratings to differ
between
the
sexes
as
a
functon
of
the
odorants
evaluated
(Gender
by
Microfragrance
TM
sample Interaction
F(49,49)= 1.64, p =0.043). The few odorants rated
more familiar
by men than by women
(e.g., coconut, root beer, tomato and honey) did
not appear to
be stereotypically
other stimuli.
总体而言,
< br>两性之间在熟悉度上的评分存在一个明显的区别,
并藉此可以将此视
为被评测气味剂的一种功能
(性别与
Microfr
agrance
TM
样品间相互影响
F
(
49,49
)
=1.64
,
p=0.043
)。
与女性相比,被男性评为熟悉度更高的气味剂为数不多(例
如,椰子,根汁汽水,番茄和
蜂蜜),这些为数不多的气味并不是那种显得刻板
“男性化”的气味,同其他刺激源相比
也并没有表现出系统性差异。
The
data
presented
in
Fig.
3
reveal
that
the
stimuli
differed
considerably
in
their
ability
to be identified. Although a number of the
odorants were correctly identified
by
most of the subjects, this was not true for the
majority. Indeed, a few were never or
only rarely correctly identified (e.g.,
pumpkin pie, honey, skunk, tomato, apple, peach,
leather, whiskey, and gingerbread).
图
3
表明,各类刺激源的被识别能力是明显不
同的。虽然许多的气味剂被大部分
的受试者正确地识别出来,
但
并不是绝大部分的气味均如此。
事实上,
一些气味
剂从来没有或只有很少被正确识别
(例如,
南瓜饼,
蜜,
臭鼬,
番茄,
苹果,
桃,
皮革,威士忌和姜饼)。
FIG. 3. Identifiability
of the 50 target microencapsulated odorants. See
text for details.
To
ascertain
whether the poorly identified
stimuli
were rated as less intense or
less
familiar
than
the
others,
as
well
as
to
determine
if
any
of
the
Microfragrance
TM
samples could be judged unsuitable for
further consideration on grounds of being too
weak, strong, unpleasant or irritating,
the means for all psychological attributes were
carefully
evaluated
for
all
50
odorants
(Fig.
2).
In
general,
those
stimuli
that
were
most
difficult
to
identify
(Fig.
3)
were
the
ones
rated
as
least
familiar
(Fig.
4).
Furthermore, the stimuli
with the lowest identifiability scores tended to
be given the
lowest
relative
intensity
scores,
despite
the
fact
that
the
intensity
ratings
and
subjective reports of
the subjects revealed them to be moderately
strong. Thus, of the
15 stimuli with
the lowest identifiability scores, approximately
two-thirds fell below
the median
intensity of the entire 50 odorants and nearly
half fell within the lowest
quartile of
the intensity ratings.
为了判定被识别情况不佳的刺激源是
否比其他刺激源具有更低的强烈度或熟悉
度,同时也为判定是否有任何
< br>Microfragrance
TM
< br>样品可能由于太弱、强烈、非愉
悦或刺激而并不适合进一步考虑,
对于所有
50
种气味剂的所有心理参数的得分均
进行了仔细的评估(
Fig.
2
)。总体来说,这些最难于识别的刺激源(
Fig.3
)均
是在熟悉度上得分最低的
(
F
ig.4
)
。
而且,识别度得分最低的
刺激源也是强烈度
得分相对最低,
尽管强烈度得分和受试者的主
观报告揭示出这些刺激源其实是中
度强烈度。
因而,
在识别度得分最低的
15
种刺激源中,
有大约三分之二的刺激源
位于所有
50
种气味剂的中间强度之下,
同时几乎有一半刺激源位于强烈度得分的
最低四分位数范围内。
FIG. 4. Relation between the
identifiability score obtained for the 50 odorants
(Fig. 3)
and the mean familiarity
rating given to the odorants(Fig. 2).
It
is
clear
from
Fig.
2
that
none
of
the
stimuli
could
be
eliminated
from
further
consideration on the basis of being too
weak or too strong, although a number were
suspect, as indicated above, in terms
of familiarity and identifiability. An examination
of the pleasantness ratings indicated
that slightly over half of the Microfragrance TM
samples were rated on the pleasant side
of the pleasantness/unpleasantness scale, and
that only a few were rated as markedly
unpleasant--motor oil, gasoline, natural
gas,
onion,
paint
thinner,
garlic,
and
turpentine.
With
the
exception
of
a
few
stimuli
of
questionable identifiability (e.g.,
honey, tomato), odorants rated as pleasant were
ones
associated with foodstuffs and
flowers, whereas those rated as unpleasant were
ones
associated with certain spices
(e.g., black pepper) or non-food objects (e.g,,
motor oil).
Irritation ratings greater
than 5 were present for 19 of the stimuli,
although none were
judged as being
extremely irritating. In general, odors rated as
unpleasant tended to
receive higher
irritation ratings than those rated as pleasant.
Most stimuli were rated
as neither warm
nor cool, although the exceptions are noteworthy.
Stimuli receiving
the largest cool
ratings (e.g., mint, menthol) and warm ratings
(e.g., cinnamon, clove,
gasoline,
motor
oil,
onion,
smoke)
are
ones
believed,
on
other
grounds,
to
produce
trigeminal
stimulation [4,16].
由图
2
可以看出,
没有任何刺激源因为太弱或太强的原因而被排除出进一步考虑,
虽然上述获得的熟悉度和可识别性数据可疑。
对愉悦度得分的分
析表明,
超过一
半的
Microfra
grance
TM
样品在愉悦
/
p>
非愉悦范畴量表上位于愉悦一侧,
并且仅有几
种样品的得分为明显的非愉悦
——
机油、汽油、天然气、洋葱
、油漆稀释剂、大
蒜和松节油。除可识别性存疑的几种刺激源外(例如,蜂蜜,番茄),
得分为愉
悦的气味剂都是与食物和花有关联的,
而得分为非愉悦
的气味剂都是与某种特定
香料有关(例如,黑胡椒)或非食物性的物质(例如,机油)。
刺激度得分超过
5
分的刺激源有
19<
/p>
种,但并没有一个被判定为极其刺激。总体来看,相比于得分
为愉
悦的气味,
得分为非愉悦的气味更偏向于获得更高的刺激度得分。
大部分刺
激源被评分为既不暖也不冷,虽然也有明显的特例。获得最大冷得分(例如,
薄
荷,薄荷脑)和暖得分(如,肉桂,丁香,汽油,机油,洋葱,烟)的刺激源,
从其他角度来看,被认为是能产生三叉神经刺激的物质。
As indicated in Fig. 2,
both the sandpaper and pencil procedures produced
clear and
similarly-intense release of
the odorants. Based on the findings that odors
released by
sandpaper
were
judged
as
more
familiar
than
those
released
by
pencik
and
that
women rated odors released in this
fashion as stronger than did men (in accord with
the sex difference noted in other human
olfactory work) [13, 17-19], we chose to use
sandpaper
for
releasing
the
stimuli
in
subsequent
studies.
Although
several
of
the
Microfragrance TM samples were
difficult to identify, it was conceivable that
some
would
be
readily
identified
in
a
multiple-
choice
situation
where
the
subjects'
responses are cued by written
alternatives (since such a paradigm generally aids
recall)
[4]. Thus, we decided to gain
more information about these stimuli in Experiment
2
before concluding they should be
eliminated from the stimulus set.
如图
2
所示,砂纸和铅笔均产生明确而同样强烈度的气味释放。基于以下两
项实
验结果,
即用砂纸释放的气味比铅笔释放的被判定为更熟悉
以及女性判别这种方
式释放的气味比男性的判别更为强烈
(同其
他有关人类嗅觉研究发现的性别差异
是一致的)
[13
,
17-19]
,我们在随后的研究中选择使用
砂纸释放刺激源。虽然有
几个
Microfragrance
TM
样品难于识别,可以想象得到,一些气味在设置了多重选<
/p>
项条件下是很容易被识别的,
这是因为此时受试者的反应受到了书
面选项的提示
(因为这样的模式通常有助于回忆)
[4]
。因此,我们决定在实验
2
中获得更多的
p>
有关这些刺激源的信息,然后再决定是否将他们从刺激源组合中排除。
EXPERIMENT 2
实验
2
Experiment 2
had three main goals: First, to ascertain whether
the odorants with poor
identifiability
tested in Experiment 1 were more readily
identified in a multiple-choice
response
situation
where
response
alternatives
were
provided:
second,
to
use
this
information, along with the information
from Experiment 1, to eliminate stimuli front
inclusion in the fnal version of the
UPSIT which were not correctly responded to by
the
majority
of
a
large
number
of
normal
individuals;
and
third,
to
evaluate,
using
multiple regression analysis, the
relative influences of several subject variables,
both
alone and in combination, on UPS1T
scores of a rather large and heterogeneous group
of subjects.
Variables
of particular interest were
those of
gender, age and
smoking.
Although most
studies suggest that olfactory ability decreases
with age [6, 36, 52, 54],
exceptions
exist [48,51]. and one recent reviewer suggests
the issue is not yet resolved
[24].
Similarly,
general
consensus
is
not
present as
to
whether
gender
and smoking
behavior influence smell
ability. For example, some studies report a
relation between
decreased
olfactory
performance
and
smoking
behavior
[30,34,
42],
whereas
others
report no such
relation [24, 40. 54].
实验
2
p>
主要有三大目的:
第一,确认实验
1
中经测试为可识别性不强的气味剂在设置多重反应选项情形下
是否会更易于被识别,
此种情形下将提供多种反应选项。
第二,
联合利用本实验
及实验
1
的信息,从最终纳入
UPSIT
范
围的的刺激源中剔除那些被绝大多数的正
常人员不能正常应答的刺激源。
第三,
采用多元回归分析方法,
评估几种受试者
变量单独或联合对大规模的不同类型人群
UPSIT
< br>得分的相对影响。
特定的目标变
量包括性别、
年龄及吸烟。
虽然大多数研究表明,
嗅觉能力随着
年龄而逐渐下降
[6, 36, 52, 54]
,但也有例外
情况存在
[48,51]
,特别是最近的一篇综述表明该问题<
/p>
仍然并没有得到解决
[24]
。同样的,
对于性别和吸烟行为是否影响嗅觉功能也还
没有形成共识。
例如
,
有些研究报告了吸烟行为和嗅觉功能下降是有关联的
[30,
34,
42]
,然而也有其他一些研究并没有发现存在这种关
联性
[24, 40. 54]
。
Subjects
受试者
In
the
initial
segment
of
Experiment
2
(where
the
identifiability
of
the
stimuli
was
established),
1198
volunteers
were
tested.
These
individuals
were
comprised
of
(a)
participants
of
regional
health
fairs
and
public
events,
(b)
primary
and
secondary
public school
students,
(c) university
students, (d) employees
of the Hospital
of the
University of
Pennsylvania, and (e) residents of homes for the
elderly. Only persons
who reported no
smell abnormalities and who were able to correctly
identify at least
half
of
the
odorants
presented
to
them
were
included.
Seventy-three
percent
of
the
group were
White Americans and 21% Black Americans, with most
of the remainder
failing
to
indicate
their
ethnicity.
Sixty-two
percent
were
female
and
38%
male.
Eighty
percent
reported
being
current
non-smokers
and
1~
reported
being
smokers,
with the remainder
not providing this information. Although a wide
spectrum of ages
was well represented
in this group, disproportionately more of the
subjects fell within
the
younger
age
categories,
as
indicated
by
the
following
summary
statistics:
mean
age=35.24,
SD--19.21;
modal
age=lg.0;
median
age=29.29;
25th
percentile:18.88;
75
th
percentile
=50.3). In general, the average ages of the two
sexes, of the two major
ethnic
groups,
and
of
the
smokers
and
non-smokers
were
similar
(e.g.,
male
mean
age=34.36
(SD:18.36);
female
mean
age=35.79
(SD:19.72);
White
mean age=36.12
(SD=20.06);
Black
mean
age=35.22
(SD=16.82);
smoker
mean
age=38.50
(SD=16.22): ramsmoker mean age=34.57
(SD= 19.69).
在实验
2
的初始阶段(此时刺激源的可识别性已经确认),
1198
名自
愿者接受测
试。这些自愿者包括:(
a
)参与地区卫生展览会和公众活动的人员,(
b
)小学
和中学公立学校学生,
(
c
)
大学生,
(
d
< br>)
宾夕法尼亚大学附属医院的员工,
(
< br>e
)
敬老院居民。
只有那些声明
没有嗅觉异常,
并且对于提供给他们的气味剂至少能
够正确识别
一半的人员才能入选。
所有人员中,
美国白人和美国黑人的比例
分别
为
73%
和
21%
,其他人员没有说明他们的种族背景。所有人员中女性和男性的比
例分别为
62%
和
38%<
/p>
。
80%
的人报告目前没有吸烟,
19%
的人报告为吸烟,其他
人员没有提供此
信息。
虽然此组人员中覆盖的年龄范围很广,
但绝大部分受试者
仍归属于年轻的年龄段,年龄统计数据总结如下:平均年龄
=3
5.24
,标准偏差
SD=19.21
;模态年龄
=19.0,
,中间年龄
=
29.29
;第一四分位数
=18.88
,第三四分
位数
=50.3
。总体来
说,两种性别、两种主要种族背景、吸烟者和非吸烟者的平
均年龄是相近的
(例如,
男性平均年龄
=34.46
(
SD=18.36
)
)
p>
;
女性平均年龄
=35.79
(
SD=19.72
)
;<
/p>
白人平均年龄
=36.12
(
SD=20.06
)
;
黑
人平均年龄
=35.22
(
SD=16
.82
)
;
吸烟者平均年龄
=38.50
(
SD=16.22
);非吸烟者平均年龄
=34.57
(
SD=19.69
)。
In the second segment of Experiment 2
(where the relative influences of factors such
as
age,
gender,
race
and
smoking
habits
were
evaluated
using
multiple
regression
analysis), the data from most of the
aforementioned subjects and an additional number
of persons (mostly elderly) were
subjected to analysis. Although, as indicated in
the
results
seclion,
UPSIT
scores
of
1365
individuals
were
evaluated
in
the
initial
multiple
regression
studies,
data
from
26
with
apparent
anosmia
or
dysosmia
were
omitted from the data set upon which
our final regression equation was computed.
在实验
2
的第二阶段(此时有关年龄、性别、种族及
吸烟习惯等因素的相对影响
已经通过多元回归分析方法进行了评估)
,
来自上述大部分受试者和额外数量的
人员
(主要为老年人)
的数据被归集进行分析。
虽然,
p>
如结果部分所述一样,
1365
名受试者的
UPSIT
得分通过初步的多元回归分析进行了评估,
在将
26
名具有明显
失
嗅或嗅觉障碍人员的数据被排除后,计算形成了最终的回归方程式。
Odorant Presentation Format
气味呈递方式
A preliminary 5-booklet 50-item version
of the UPSIT, identical in general format to
the 40-item version, was developed and
administered in this experiment. The order of
presentation
of
the
odors
within
the
booklets
(10
odorants
per
booklet)
was
determined randomly, with the exception
that odors of similar psychological quality
(e.g.. onion, garlic) did not directly
follow one another. Although a large number of
potential descriptors were considered
for use as response alternatives [21,28], the 50
descriptors
assigned
by
the
manufacturer
to
the
Microfragrance
TM
samples
were
finally chosen, as they were relatively
unambiguous and non-esoteric. One additional
descriptor,
was
also
included
in
this
group
to
help
achieve
specific
criteria
outlined below.
一种总体上类似于
40
项版本的
50
项<
/p>
5
—手册的初步
UPSIT
版本被建立并应用于本
次实验中,
每一手册(每个手册
10
种气味剂)中的气味呈递顺序
是随机的,但
不会将心理感受类似的气味(例如,洋葱和蒜头)连续放在一起。虽然许多
潜在
的描述符可被考虑用于作为应答选项
[21,28]
,但最终选取了生产商所分配给
Microfragrance
TM
样品的
50
中描述符,
因为这些描述符是相对简单明了和易懂的。
p>
一个额外的描述符“
cola
”也被纳入进
来以帮助达到下面所列出的特定标准。
To aid in the selection of sets of
distinct descriptors for each stimulus, the names
of
all 51 descriptors were typed on
small cards. These cards were then arranged by two
female
research
technicians
spatially
on
a
table
top
with
the
goal
of
making
the
distances between them
reasonably proportional to the psychological
similarity of the
odor items. For
example, the onion and garlic labels were placed
close to one another,
whereas the
chocolate and gasoline labels were placed apart
from one another and at
differing
distances
from
those
of
onion
and
garlic.
The
general
arrangement
agreed
upon by the two
technicians was subsequently evaluated by one of
the experimenters
(RLD), who concurred
with the general arrangement. This simple
procedure resulted
in a two-dimensional
space from which the three
each odorant
so as to maximize their distinctiveness from one
another as well as from
the target
stimulus.
为了帮助所有的刺激源选取一套独特的描述符,
< br>所有
51
个描述符的名字被打印在
小卡片上。
然后由两名女性研究技术员将这些卡片排列于桌面上,
并确保心理感
受相似的气味剂之间保持合适的距离。
例如,
洋葱和蒜头标签相互靠近,
而巧克
力和
汽油标签则相互远离,
不同于洋葱和蒜头两者之间的距离。
总体
摆放一旦获
得两名技术员的认可后即由其中一名实验人员进行评估并获得其认同
。这种
简单的程序产生了一种二维空间,
p>
基于此二维空间,
可以为每一气味剂选取三种
干扰项,以便将各个气味剂之间及同目标气味剂之间的差异最大化。
In addition to selecting
response alternatives as distinct from one another
as possible,
we
sought
to
use
each
of
the
descriptors
equally
often
and
approximately
the
same
number of times in the
a, b, c and d response category positions.
Unfortunately, it was
not possible to
achieve all of these aims simultaneously.
Nonetheless, we approached
this ideal.
Thus, 36 of the 51 descriptors appeared four times
apiece, eight three times
apiece,
five
five
times
apiece,
one
six
times,
and
another
once.
No
descriptor
ever
appeared more than twice in any of the
response category positions (i.e., a, b, c or d).
除了选取尽可能相互不同的应答选项,
我们也努力试图以同样频
率和同样次数在
a
,
b
,
c
,
d
选项中使用每一个描述符。当然,并不可能同时达到所有这些目标,
这只是理想状态
下的一种目标。
因此,
51
种中的描述
符有
36
种各自出现了
4
次,
8
种各出现了
3
次,
5
种各出现了
5
次,
1
种出现
6
次,另一种出现
1
次。没有任何描述
符在任何应答选项位置(例如,
a
,<
/p>
b
,
c
或
d
)出现超过两次。
Procedures
流程
Each subject
was
instructed to
complete the five booklets
in order. A response was
required for each item, even if no odor
was perceived (i.e., the test was forced-choice).
In
most
cases
subjects
were
observed
while
taking
the
test
and,
in
a
few
instances
where the instructions were not
understood, they were clarified. Following
completion
of the tests, they were
checked to insure that all items had been
answered.
In cases
where an
item was skipped, a subject was instructed to
complete the missing item.
每位受试者都被要求依次完
成
5
本手册。对于每一项目均需选择一个应答选项,
即使没有感知到任何气味(此测试是属于强制性的选择)。在大多数情况下,受
试者在进行测试时都会有人在观察,
当受试者对试验指南没有完全明白时会对其
p>
进行解释。
一旦完成了测试后,
将会进行核
查以确保所有项目均已回答完毕。
如
果出现有项目被跳过,受试
者将会被要求完成遗漏的项目。
Although the test was self-administered
in most cases, an experimenter administered
the test in situations where the person
being tested was incapacitated or quite old. In
these cases, the response alternatives
were read aloud to the subject both before and
during the sniffing of each stimulus,
which was held directly underneath the subject's
nose.
虽然此次测试在大多情况下是自行完成的,
但在一些特殊情况下,
如受试者由于
身体不便
或年老的原因不能自行完成测试,
将会有实验人员协助完成测试。
在这
些情况下,
应答选项将会在每一种刺激源闻之前和闻的过
程中被大声的读给受试
者听,同时会将刺激源置于受试者的鼻子底下。
< br>
RESULTS
结果
Odor
ldentifiability
气味识别
Even
though
the
responses
were
cued
by
printed
alternatives,
a
number
of
stimuli
were still poorly identified by the
majority of the subjects (Fig. 5). For example,
less
than
50%
of
the
subjects
correctly
identified
the
so-called
honey
and
whiskey
fragrances, and less than 75% were able
to correctly identify the skunk, pumpkin pie,
tomato, chili, black pepper, and apple
odors. In general, the stimuli that were poorly
identified in this study were the same
ones that were poorly identified in Experiment 1.
For example, of the 15 least
identifiable odorants in Experiment 1 (Fig. '3),
13 were
amongst the 15 most poorly
identified odorants in Experiment 2 (Fig. 5).
Despite this
fact,
however,
the
inclusion
of
response
alternatives
did
result
in
a
higher
relative
identifiability
of
most
of
the
stimuli,
as
was
expected
from
previous
work
[4].
In
addition,
as was also expected from previous studies [4, 13,
17-19], women performed
better than men
in correctly identifying most of the odorants
(Fig. 5).
虽然所有的应答均以打印选项来提示,
许
多刺激源仍然并没被大部分的受试者识
别出来
(
Fig.5
)
。
例如,
只有不到
50%
的受试者正确识别出了名为<
/p>
“蜂蜜”
和
“威
士忌”的气味,不到
75%
的受试者正确识别出了臭鼬、南瓜饼
、番茄、辣椒、黑
胡椒和苹果的气味。
总体来说,
在本次实验中,
识别情况不理想的气味基本上和
实验
1
中识别情况不理想的气味相同。例如,在实验
1
中被最少人识别出来的
15
种气味剂
(
Fig. 3
)
,
就有
13
种位列实验<
/p>
2
中被最少人识别出来的
15
种气味
(
Fig.5
)<
/p>
。
尽管如此,
然而,
应答选项的设置的确导致绝大部分刺激源具有了相对更高的可
识别性,这和之前的研
究结果是一致的
[4]
。除此之外,也跟之前的研究结果一
p>
样
[4,13,17-19]
,在正确识别
绝大多数的气味方面女性比男性表现更佳(
Fig.5
)。
p>
Based on these
data and the identifiability data of Experiment 1,
the following stimuli
were
eliminated
from
inclusion
in
the
40-item
UPSIT:
apple,
black
pepper,
chili,
honey,
musk,
pumpkin
pie,
skunk,
tomato
and
whiskey.
In
addition,
garlic
was
eliminated from the
final test due, in part, to its psychological and
chemical similarity
to onion.
基于这些数据和实验
1
的可识别性数据,以下刺激源被从<
/p>
40
项
UPSIT
组合物中剔
除:苹果,黑胡椒,辣椒,蜂蜜,麝香,南瓜饼,臭鼬,番茄和威士忌。此
外,
蒜头也被从最终的测试组合物中排除,
这部分原因是因为蒜
头的心理感受和化学
特性都类似于洋葱。
Relationship of Age,
Gender, Race and Smoking Habits to UPSIT Scores
年龄、性别、种族和吸烟习惯与<
/p>
UPSIT
得分的关系
All subsequent studies were performed
using only the 40-stimuli included in the final
version of the UPSIT. To examine to
what extent a number of demographic factors
account
for
the
variability
observed
in
the
UPSIT
scores,
we
performed
a
series
of
multiple regression analyses on data
from 1339 to 1365 subjects, depending upon the
information available (i.e., missing
data
for some variables necessitated
using fewer
subjects).
Because we wished to consider only subjects with
reasonably normal smell
function,
persons
with
UPSIT
scores
less
than
20
were
excluded
from
most
of
our
analyses
(inclusion
of
such
individuals
has
the
potential
for
greatly
increasing
the
variance and producing a
spurious R ~ value). The fitted regression
equation (for I339
subjects)
which
excluded
such
persons
and
included
only
variables
statistically
significant at the 0.05 level (F-test)
was as follows:
Y=33.399+1.055X
< br>1
+0.217X
2
-0.00
3 X
2
2
-0.489 X
3
-1.008X
4
-
1.040X
5
-
2.172X
6
+ e,
where: X
1
= 1(0)
if the subject is female (male);
X
2
= age of
subject in years;
X
3
= 1(0) if the
subject does (does not) currently smoke;
X
4
= 1(0) if the
subject is (is not) nonwhite;
X
5
= 1(0) if the
subject does (does not) report a smell problem;
X
6
=
1(0)
if
the
subject
does
(does
not)
belong
to
an
elderly
sub-file
(i.e.,
persons primarily in old-age homes who
are over 65 years of age);
e = error
term.
The
value
of
R
2
for
this
equation
was
0.411
(SE=3.318).
The
standard
errors
of
estimate for the seven
variables were: X,=0.188; X2=0.023; X22=0.0003;
Xa=0.238;
X4=0.222; X~=0.302; and
Xa=0.525.
所有接下来的研究都仅仅使用被包含在最终版本
< br>UPSIT
中的
40
种刺激源来
实施。
为了探明许多人口统计学因素在多大程度上会影响所观察到的
UPSIT
得分变化,
我们对来自
1339
—
1365
名受试者的数据进
行了一系列的多元回归分析,
具体受试
者数量依赖于数据的可用
性
(例如,
对于一些有必要采用少一点受试者的变量而
导致数据的遗漏)。因为我们希望仅仅考虑具有合理正常的嗅觉功能的受试者,
< br>UPSIT
得分少于
20
分的人
员都被排除在我们的大部分分析之外
(如果这类人员纳
入到分析
中来,
可能会极大的增大方差和产生虚假的
R
< br>2
值)
。
排除了此类人员之
p>
后所得拟合回归方程
(对应
1339
名受试者)
如下,
此方程仅仅包含在
0.05
水平
(
F-<
/p>
测试)具有统计学显著意义的变量。
Y=33.399+1.055X
1
+ 0.217X
2
-
0.003X
2
2
-
0.489X
3
-1.008X
4
-1.040X
5
-
2.172X
6
+ e
公式中:
X
1
= 1
(
0
),如果受试者是女性(男性)
<
/p>
X
2
=
受试者年
龄(以年记)
X
3
= 1
(
0
),如果受试者目前在吸烟(非吸烟)
X
4
= 1
(
0
),如果受试者是(不是)非白人
X
5
= 1
(
0
),如果受试者报告(没有报告)了嗅觉问题
X
6
= 1
(
0
),如果受试者是属于(不属于)老年阶段(
例如,主要为来自养
老院的人员,均超过
65
< br>岁)
e=
误差项
该方程式中,
R
2
值是
0.411
(
SE=3.318
)
。
对于七个变量估测的标准误差分别是:
X
1
=0.188
,
X
2
=0.023
,
X
2
2
=0.0003
,
X
3
=0.238
< br>,
X
4
=0.222
,
X
5
=0.302
,
X
6
=0.525
。
We
carried
out
several
regression
analyses
with
an
additional
dummy
variable
to
permit inclusion of the twenty-six
subjects who reported a smell problem and scored
less than 20 correct. A fitted model
comparable to the one above (with an n of 1365)
yielded an R 2 value of 0.608, although
the estimate of the standard deviation of the
error term (SE) was 3.345, nearly
identical to the aforementioned model's value.
同时,
我们也对一个虚拟变量进行了几次回归分析,
< br>以便将
26
名报告有嗅觉问题
并
且得分少于
20
分的受试者纳入进来。
相比于以上方程,
一个拟合的模型
(对于
1365
名受试者)
获得的
R
2
值为
0.608
,
然而误差项
(
SE
)
的标准偏差估测值是
3.345
,
几乎同上述方程式的值等同。
FIG. 5. Percent of individuals
correctly identifying the 50 microencapsulated
odorants
in
the
4-alternative
forced-choice
response
paradigm.
Note
that
women
performed
better than men on
this task for 44 of the stimuli.
These
data
indicate
that
gender,
age,
ethnic
background,
and
smoking
habits
all
significantly correlate, in varying
degree, with scores on the UPSIT. The relation of
gender
and
age
to
median
UPSIT
values
is
clearly
depicted
in
a
recent
analysis
of
1447 subjects (Fig. 6).
Note that women evidenced higher average UPSIT
scores than
did men across nearly all
age groups. A systematic decrease in test
performance was
present for both sexes
beginning in the seventh decade of life and
continuing until the
tenth decade, with
a more marked decrease occurring in men.
这些数据表明,性别、年龄、种族背景和吸烟习惯,都在不同程度上同
UPSIT<
/p>
得