VaughanSellersWolpaw(2012)

Full text

325
86
85
84
83
82
81
80
79
78
77
76
75
74
73
72
71
70
69
68
67
66
65
64
63
62
61
60
59
58
57
56
55
54
53
52
51
50
49
48
47
46
45
44
43
42
41
40
39
38
37
36
35
34
33
32
31
30
29
28
27
26
25
24
23
22
21
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
P
revious chapters in this book have discussed the techni-
cal principles and methods of BCI technology.  ese
chapters show that, despite their current limitations,
BCIs are fast becoming e ective communication and control
devices. However, the rapid growth of this research and its
remarkable progress are still con ned almost entirely to the
cosseted environments of a multitude of laboratories through-
out the world. Furthermore, most BCI experiments have been
and continue to be conducted in able-bodied humans or ani-
mals rather than in the severely disabled people for whom this
new technology is primarily intended.
Certainly, there are compelling theoretical and practical
reasons for this overwhelming focus on laboratory studies in
normal subjects: labs provide the strictly controlled environ-
ments and expert oversight conducive to the development and
optimization of new technology; and able-bodied populations
are more available and avoid the additional variables intro-
duced by disease and injury that may vary widely across
individuals.
Nevertheless, this focus leaves a major research gap that
must be addressed if BCIs are to ful ll their primary purpose
and justify the considerable support that their development
receives from governments and other funding entities.  at is,
the BCIs that work well in the laboratory need to be shown to
work well in real life, to provide people with disabilities new
communication and other capabilities that improve their daily
lives.
In some ways, this essential task is considerably more com-
plicated and more demanding than the laboratory research that
produces a BCI system.  at original research has a single aim:
to design and optimize a BCI that provides reliable and accurate
communication or control in a carefully controlled and closely
monitored laboratory setting. In contrast, research that seeks to
establish the real-life usefulness of a BCI system has four di er-
ent aims.  ey may be stated as a set of four questions:
Can the BCI design be implemented in a form
•
suitable for long-term independent use?
Who are the people who need the BCI system,
•
and can they use it?
Can their home environments support their use
•
of the BCI, and do they actually use it?
Does the BCI improve their lives?
•
 is chapter addresses each of these questions in turn.
It considers the steps involved in answering each and the
potential problems that must be overcome. Since the present
peer-reviewed literature lacks any formal multisubject studies
that address these questions (and indeed has few reports of any
kind that are directly relevant to these questions), the discus-
sion necessarily relies heavily on the authors’ experience to
date, which is primarily with a noninvasive EEG P300-based
BCI system (see chapter 12 in this volume). Nevertheless, the
chapter’s overall intent is to provide information and insight
that would apply to any e ort to take any BCI system out of the
lab and validate its e ectiveness in the everyday lives of people
with disabilities.
CAN THE BCI DESIGN BE IMPLEMENTED
IN A FORM SUITABLE FOR LONG-TERM
INDEPENDENT USE?
For some BCIs, this  rst question is readily answered in the
negative. For example, the expense, size, and complexity of
fMRI-based or MEG-based BCI systems con ne them to labo-
ratory settings, at least for the foreseeable future (Bradshaw
et al. 2001 ; Buch et al. 2008 ; Cohen 1972 ; Kaiser et al. 2005 ; Lee
et al. 2009 ; Mellinger et al. 2007 ; Tecchio et al. 2007 ; van Gerven
and Jensen 2009 ). BCIs that rely on implanted devices (e.g.,
electrocortigraphy [ECoG], local  eld potentials [LFPs], or
single units) have demonstrated impressive capacity both in
animals and in humans.  ese BCIs face the same safety
requirements as any device for clinical use, and, in addition,
they must demonstrate that they are su ciently reliable and
e ective to warrant human implantation (Donoghue 2008 ). At
present, BCIs based on EEG (and possibly also those based on
functional near-infrared spectroscopy [fNIRS]) are the best
candidates for independent use (Bauernfeind et al. 2008 ; Coyle
et al. 2007 ; Naito et al. 2007 ). Even so, their transition from the
laboratory to the home, and to long-term everyday use, requires
substantial recon guration of their components and consider-
ation of issues that do not generally arise in the laboratory.
Any BCI system deployed for independent use must be
safe to operate in the home environment without on-site tech-
nical support. Components should be few, small, portable, and
relatively inexpensive; and the connections between them
should be minimized (e.g., by use of telemetry) and extremely
robust.  ey must be packaged in sturdy and con gurable
20 CLINICAL EVALUATION OF BCIS
THERESA M. VAUGHAN , ERIC W. SELLERS , AND JONATHAN R. WOLPAW
20-Wolpaw-20.indd 325
20-Wolpaw-20.indd 325
8/10/2011 5:14:40 PM
8/10/2011 5:14:40 PM
OUP UNCORRECTED PROOF – FIRST-PROOF, 11/08/2011, GLYPH

326 | BRAIN–COMPUTER INTERFACES
47
46
45
44
43
42
41
40
39
38
37
36
35
34
33
32
31
30
29
28
27
26
25
24
23
22
21
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
housing to provide  exible setup and easy storage and must be
able to withstand potentially rough handling over many
months. Ideally, the ampli ers should be insensitive to the
many sources of electromagnetic noise present in home set-
tings, and the electrodes and their mounting (e.g., for EEG, the
electrode cap) should be capable of functioning safely and
e ectively for many hours per day over months without main-
tenance or replacement.  e so ware should be easy to use and
thoroughly tested (i.e., impervious to BCI user or caregiver
error). Before attempting to take a BCI system out of the labo-
ratory, investigators should meet these requirements to the
greatest extent possible. At the same time, they should recog-
nize that further changes are likely to be needed when the BCI
is actually deployed in the home environment. In this regard
the principles of modularity in the so ware (e.g., Schalk et al.
2004 ) and in the hardware (e.g., Cincotti et al. 2008 ) can expe-
dite the implementation of improvements and upgrades, and
the tackling of unexpected failures. Figure 20–1 A shows the
current version of the P300-based BCI home system developed
at the Wadsworth Center of the New York State Department of
Health (Albany, NY); and  gure 20–1 B shows a compact trav-
eling unit for evaluating this system’s suitability for potential
users who are homebound.
Figure 20–2 A shows the Wadsworth BCI home system in
operation.  is system has now been used by seven severely
disabled people in their homes over months and years. It is
managed by the caregivers in the users’ homes, with internet
oversight from the Wadsworth BCI laboratory and occasional
home visits by technical personnel from the lab.  e foreground
of  gure 20–2A shows the crowded environment of the user’s
room. It is typical of the environments of people with severe
disabilities.
WHO ARE THE PEOPLE WHO NEED THE
BCI, AND CAN THEY USE IT?
Present-day BCIs have relatively modest capabilities.  us, the
communication and control applications they can provide are
likely to be of signi cant value only to people with extremely
severe disabilities that prevent them from using conventional
assistive technologies (see chapter 11). Over the past decade a
number of studies have begun to explore the BCI capacities of
people severely disabled by disorders such as ALS or high-level
spinal cord injury (e.g., Bai et al. 2010 ; Bai et al. 2010 ; Birbaumer
et al. 1999 ; Conradi et al. 2009 ; Farwell and Donchin 1988 ;
Hochberg et al. 2006 ; Ho mann et al. 2008 ; Ikegami et al. 2011 ;
Kauhanen et al. 2007 ; Kennedy and Bakay 1998 ; Kubler
et al. 2001 ; Kubler et al. 2005a ; Kubler et al. 2009 ; McFarland
et al. 2010 ; Miner et al. 1998 ; Mugler et al. 2010 ; Muller-Putz
Figure 20.1 (A) The current Wadsworth P300-based BCI home system. The components include a laptop computer, an eight-channel EEG amplifi er
(Guger Technologies,), an electrode cap (Electro-Cap International,), a 20” monitor, and connecting cables. (B) A compact traveling BCI evaluation unit
designed for easy setup, breakdown, and storage of all necessary hardware and supplies.
Figure 20.2 (A) A person severely disabled by amyotrophic lateral sclerosis (ALS) using the Wadsworth brain-computer interface (BCI) system in his home. He wears
a modifi ed eight-channel electrode cap. (B) Monitor display used by caregiver to check electrode impedance. Red dots are the locations of the eight recording
electrodes. When all the locations become green, electrode impedance is suffi ciently low, and the caregiver can begin the BCI session.
20-Wolpaw-20.indd 326
20-Wolpaw-20.indd 326
8/10/2011 5:14:40 PM
8/10/2011 5:14:40 PM
OUP UNCORRECTED PROOF – FIRST-PROOF, 11/08/2011, GLYPH

CHAPTER 20. CLINICAL EVALUATION OF BCIS | 327
103
102
101
100
99
98
97
96
95
94
93
92
91
90
89
88
87
86
85
84
83
82
81
80
79
78
77
76
75
74
73
72
71
70
69
68
67
66
65
64
63
62
61
60
59
58
57
56
55
54
53
52
51
50
49
48
47
46
45
44
43
42
41
40
39
38
37
36
35
34
33
32
31
30
29
28
27
26
25
24
23
22
21
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
et al. 2005 ; Nijboer et al. 2008 ; Pfurtscheller et al. 2000 ; Piccione
et al. 2006 ; Pires et al. 2011 ; Sellers and Donchin 2006 ;
Sellers et al. 2010 ; Silvoni et al. 2009 ; Townsend et al. 2010 ;).
Although some subjects have been studied in their home envi-
ronments, most of this work has generally consisted only of
limited sessions with the experimenters closely overseeing
BCI operation. Nevertheless, the results to date are encourag-
ing in that they indicate that many people with severe dis-
abilities can use BCIs that could in theory help them in their
daily lives.
 ese individuals are usually home-bound (or institution-
bound) and attended by caregivers 24 hours per day (Albert
et al. 2009 ).  ey comprise the target user population for the
BCIs that are available now or likely to be available within the
next decade. How does a BCI researcher  nd good subjects for
studies testing the e ectiveness and utility of BCI home use for
people with severe disabilities? And how does he or she proceed
with these subjects once they are identi ed?
DEFINING THE POPULATION OF
PROSPECTIVE BCI HOME USERS
As in most clinical studies, subjects are selected according to a
speci c set of criteria. For the user population described above,
the basic inclusion criteria would be:
Little or no useful voluntary muscle control
•
(e.g., people with late-stage ALS, muscular
dystrophy, severe Guillain-Barré syndrome,
brainstem stroke, severe cerebral palsy,
high-level spinal cord injury, or a variety
of other severe neuromuscular disorders).
(For people with ALS or other progressive
diseases, this criterion might be extended to
include those who have not yet reached this
level of disability but can be expected to do
so eventually.)
Conventional assistive (i.e., muscle-based)
•
communication devices (e.g., eye-gaze systems,
EMG switches) are not adequate for their
needs: they may be entirely unable to use these
devices; their control may be inconsistent or
they may fatigue quickly; they may not like the
devices; or they may desire the additional
communication and control capabilities that a
BCI could provide.
Medically stable, with the intent, and a
•
reasonable expectation, of living for at least one
year. If they have ALS, they have already begun
arti cial ventilation or have decided to do so
when it becomes necessary.
Able to follow spoken or written directions.
•
Absence of any other impairment that would •
prevent BCI usage (e.g., extremely poor vision
would prevent use of a BCI that uses visual
stimuli).
Stable living environment.
•
Reliable caregivers (family members and/or •
professionals) possessing or capable of
acquiring basic computer skills and enthusiastic
about supporting the subject’s BCI usage.
Subject and caregivers able and willing to
•
provide informed consent and clearly
enthusiastic about participating in a research
study that may have no lasting direct bene t to
them (Vaughan et al. 2006 ).
Given the wide variety of disorders that can cause severe
motor disability, the complexity of the disabilities they cause,
and other variables associated with these disorders (e.g., medi-
cation, other medical problems), it is o en di cult to deter-
mine whether a particular person satis es these criteria
(Kuebler et al. 2006). For example, aphasia, which occurs in
association with over 25 % of strokes, can interfere with the
ability to understand instructions about how to use the BCI
and/or with formulation of messages to be communicated with
it (Pederson et al. 1995 ; Wade et al. 1986 ). On the other hand,
a right or le hemianopsia (i.e., loss of the right or le visual
 eld) produced by stroke would probably not interfere with
BCI use if the screen is positioned in the remaining visual  eld.
Since many prospective BCI users are older adults with ALS or
strokes, age-related visual impairments (e.g., macular degen-
eration, glaucoma, and cataracts [Strei 1967 ]) might also
a ect BCI capability.
Appropriate assessment questions (e.g., can the person
read text on a screen?) or a standard measure of visual acuity
(e.g., Snellen test [Tucker and Charman 1975 ]) may evaluate
this visual issue. Another relevant factor includes current med-
ications (e.g., sedatives) that may interfere with brain function
or a ect the EEG (Towler et al. 1962 ). Cognitive impairments
(which occur in up to 40 % of people with ALS [Woolley et al.
2010; Volpato et al. 2010 ]) and depression may also interfere
with BCI use. Although the recent literature indicates that
people with advanced ALS generally rate their quality of life as
quite high, moderate depression is o en present (Gauthier
et al. 2007 ; Chio et al. 2004 ; Robbins et al. 2001 ; Simmons et al.
2006 ; Kubler et al. 2005b ). As in other therapeutic endeavors
(Kirchho and Kehl 2007 ) (as well as in most life endeavors),
mood can a ect motivation and play a signi cant role in BCI
e ectiveness (Kleih et al. 2010 ).
RECRUITING PARTICIPANTS FOR BCI
HOME-USER STUDIES
Subject recruitment is a key part of any clinical study and o en
presents signi cant di culties (e.g., Bedlack et al. 2010 ).
Recruiting and retaining individuals who have entered the late
stages of a progressive neurological disease can be particularly
challenging (Shields et al. 2010 ). Hospitals, regional clinics,
20-Wolpaw-20.indd 327
20-Wolpaw-20.indd 327
8/10/2011 5:14:41 PM
8/10/2011 5:14:41 PM
OUP UNCORRECTED PROOF – FIRST-PROOF, 11/08/2011, GLYPH

328 | BRAIN–COMPUTER INTERFACES
108
107
106
105
104
103
102
101
100
99
98
97
96
95
94
93
92
91
90
89
88
87
86
85
84
83
82
81
80
79
78
77
76
75
74
73
72
71
70
69
68
67
66
65
64
63
62
61
60
59
58
57
56
55
54
53
52
51
50
49
48
47
46
45
44
43
42
41
40
39
38
37
36
35
34
33
32
31
30
29
28
27
26
25
24
23
22
21
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
and medical specialists are traditional sources of subject refer-
rals. However, many potential BCI home users no longer attend
a clinic regularly or participate in routine rehabilitation ser-
vices, and they may not be under the continuing care of medi-
cal specialists. On the other hand, many of these individuals
are enrolled in programs that provide assistive technology
(AT) for seating, mobility, and communication needs (Cotterell
2008 ).  us, subject recruitment is o en accomplished by con-
tacting speech/language pathologists and/or physical thera-
pists. Home-care physicians, rehabilitation hospitals, visiting
nurse services, and hospice providers can also be sources of
potential BCI home users. Local school districts frequently
have information on programs that serve people with extreme
physical challenges. Finally, certain registries of patient popu-
lations can be useful in recruiting a clinical study cohort (e.g.,
the national registry of veterans with ALS developed by the
Veterans Administration, National ALS Registry Home Page;
Allen et al. 2008 ; Lancet Neurology Editorial 2009 ]). Such reg-
istries can expand the number of potential contacts well beyond
the immediate geographic region. Registries vary in the cur-
rency of their information and in the steps required to use
them in subject recruitment (e.g., Registry board approval,
local IRB oversight).
Whether a particular individual meets the inclusion crite-
ria de ned above can normally be determined from interviews
with caregivers, medical personnel, and/or family members.
 us, in most instances, people who do not meet the criteria
can be identi ed and excluded without actually testing them
with the BCI.  is can substantially reduce the time and e ort
the research group invests in testing people who do not turn
out to be appropriate for the study. It may also substantially
reduce the possibility that exclusion might greatly disappoint a
p r o s p e c t i v e s u b j e c t .
OBTAINING INFORMED CONSENT
 e extremely disabled people who could bene t from current
BCIs generally lack understandable speech. In many cases their
communication depends entirely on subtle movements of the
face, especially small movements of the eyes (Neumann and
Kubler 2003 ).  us, it may be di cult to obtain the subject’s
informed consent for participation in a BCI study. Nevertheless,
individuals who retain a clear capacity to control such simple
movements and to thereby communicate (e.g., via a letterboard
held by a caregiver) can provide informed consent, although
the process may require considerable time and e ort for all
concerned. Furthermore, for studies that are found to pose no
signi cant risk (e.g., most noninvasive BCI studies), subjects
may participate by providing informed assent (Black et al.
2010 ). Informed assent requires only that they be able to answer
yes/no questions. Unlike informed consent, it does not require
that they be able to ask questions.
For people in whom the capacity to provide informed con-
sent (or assent) is uncertain, many locales have established
procedures for permitting close relatives to act on behalf of
an incapacitated person to provide informed consent for
participation in a clinical trial. Although such surrogate
approvals may be relatively straightforward for noninvasive
minimal-risk BCI systems, they become more problematic for
invasive BCI systems, which may entail signi cant risks
(including possible discomfort) (see discussion in chapter 24).
For people with progressive diseases such as ALS, informed
consent may be obtained (and BCI use might be initiated)
during earlier stages of the disease when adequate communi-
cation capacity is still present. (Early BCI use may also facili-
tate the transition to extensive BCI use when conventional
communication is no longer possible.)
DETERMINING WHETHER A POTENTIAL
STUDY SUBJECT CAN USE THE BCI
For each person who has met the inclusion criteria and pro-
vided informed consent (or assent), the next step is an evalua-
tion of his or her ability to use the BCI.  is evaluation
represents a Go/No-Go decision for participation in the clini-
cal study. In work to date by the Wadsworth BCI research
group using a P300-based BCI, this evaluation has consisted of
two or three 1–2 hour sessions. During each of these sessions
the subject performs a cued letter-selection task referred to as
copy spelling (Birbaumer et al. 1999 ).  e goal is to collect suf-
 cient data to parameterize the BCI so that henceforth the user
can then it to communicate intent (e.g., to spell freely, select
icons, etc.). In most cases, as few as 21 copy-spelling selections
(i.e., trials) are su cient to parameterize the system (McCane
et al. 2009 ). With the standard 6 × 6 P300 matrix (for which
chance accuracy is 2.8 % ), accuracy of > 70 % is generally consid-
ered adequate for e ective communication (Sellers et al. 2006 ).
McCane et al. ( 2009 ) used interviews to identify 25 people
with ALS who appeared to be good candidates for use of a
P300-based BCI. In subsequent testing with the BCI system, 17
of the 25 candidates (68 % ) achieved the requisite accuracy of
> 70 % and were thus judged able to use the BCI. It is worth
noting here that there was no correlation between the subjects’
BCI accuracy and their disability level as measured with the
ALS functional rating scale. For the remaining eight people
accuracy was <40 % . Seven of these people had visual problems
(e.g., ptosis, nystagmus, diplopia) that interfered with BCI use.
(Such problems are common in people with late-stage ALS
[Mizutani et al. 1990 ; Pinto and de Carvalho 2008 ]).  ese
data further emphasize the importance of gathering relevant
information prior to BCI testing.
 e evaluation of a person’s capacity to use the BCI may be
particularly di cult with people who lack a clearly reliable
means of basic communication (e.g., an eyeblink or muscle
twitch). If an individual does not have an obvious and rela-
tively fast way to ask and answer questions, the only way to
know that he or she has understood the instructions is for the
person to communicate using the BCI, and this requires and
assumes that the BCI itself is working properly.  e di cult
issue of BCI use by people who lack any muscle-based com-
munication (i.e., are completely locked in) is addressed more
fully in chapters 11 and 19 of this volume.
20-Wolpaw-20.indd 328
20-Wolpaw-20.indd 328
8/10/2011 5:14:41 PM
8/10/2011 5:14:41 PM
OUP UNCORRECTED PROOF – FIRST-PROOF, 11/08/2011, GLYPH

CHAPTER 20. CLINICAL EVALUATION OF BCIS | 329
106
105
104
103
102
101
100
99
98
97
96
95
94
93
92
91
90
89
88
87
86
85
84
83
82
81
80
79
78
77
76
75
74
73
72
71
70
69
68
67
66
65
64
63
62
61
60
59
58
57
56
55
54
53
52
51
50
49
48
47
46
45
44
43
42
41
40
39
38
37
36
35
34
33
32
31
30
29
28
27
26
25
24
23
22
21
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
CAN THE HOME ENVIRONMENT SUPPORT
BCI USE, AND IS THE BCI ACTUALLY
USED?
ASSESSING THE ENVIRONMENT AND THE
CAREGIVERS
Successful home use of current BCI systems require a home
environment that can support their use. Home environment
assessment can be accomplished  rst by appropriate questions
during telephone interviews and then during the initial BCI
evaluation sessions. Home assessment includes evaluation of
not only the physical environment but also the level of interest
and ability of the users and their caregivers.  e immediate
environments of people with severe disabilities are o en
crowded with much essential equipment, including ventilators,
mechanical beds, and wheelchairs.  us, the placing of the BCI
system and the positioning of the prospective user may be
challenging, and signi cant sources of electrical noise and
intermittent artifacts may be present.  ese factors, the di -
culties they present, and the prospects for overcoming them
can be initially assessed in the  rst home visits.  ese visits are
also an opportunity to assess, at least in an informal fashion,
the technical skills, learning capacities, interest, and motiva-
tion of the caregivers who will need to support BCI use.
Without capable and motivated caregivers, long-term BCI
home usage is not possible (Wilkins et al. 2009 ).
For subjects who have an adequate home environment and
are able to use the BCI, the next step is to tell the subject and
the caregivers who will support and oversee BCI use about the
BCI applications available and about the time, e ort, and spe-
ci c tasks involved in BCI use.  is will allow the level of moti-
vation of both the subject and the caregivers to be further
assessed. If they are motivated, a plan may then be formulated
incorporating the purposes for which the user wants to use the
BCI. For all users, particularly those who still retain some
capacity for conventional (i.e., neuromuscular) communica-
tion, this planning step should involve both the user and the
caregivers to the greatest extent possible. As described in chap-
ters 11 and 19, the participants’ involvement is a key factor in
the success of testing new and/or old BCI applications. If the
subjects and their caregivers are motivated and a good usage
plan has been de ned, the study can then move on to deter-
mine whether the person actually uses the BCI in daily life.
INITIATING AND EVALUATING BCI HOME USE
Initiation and evaluation of BCI home use includes  ve primary
tasks:
C o n  guring the BCI to satisfy the needs and
•
preferences of the user
Placing the BCI in the home
•
Training the subject to use the BCI applications •
and the caregivers to support BCI use
Providing ongoing technical support as needed
•
Measuring the extent, nature, and success of •
B C I u s a g e .
CONFIGURING THE BCI FOR THE USER
Before home use begins, the BCI should be con gured for the
individual user. For example, in the standard P300-based BCI,
the numbers and sizes of the matrix items, as well as their
brightness and  ash-rate, can generally be adjusted according
to the abilities and preferences of the user. Careful attention to
each user’s abilities and preferences is essential. Although
speed is generally considered important in communication, it
may or may not be of paramount importance to a user who has
little or no remaining useful motor function (Millán et al.
2010 ). For these individuals, the restoration of some measure
of independent communication may be more important than
speed.  ey may prefer slower but more accurate output to
faster but less accurate output. Indeed, one person severely
disabled by ALS who uses a P300-based BCI in his daily life
chooses to have a 9-sec pause inserted a er each selection and
thus communicates at a rate considerably slower than the max-
imum rate the BCI could provide (Sellers et al. 2010 ). When
they become available for home use, BCI systems that use audi-
tory stimuli or combined visual/auditory stimuli may be most
appropriate for people who lack su cient visual function
(Farquhar et al. 2008 ; Hill 2005 ; Guo et al. 2010 ; Hinterberger
et al. 2004 ; Klobassa et al. 2009 ; Sellers and Donchin 2006 ;
Nijboer et al. 2008 ; Furdea et al. 2009 ; Kanoh et al. 2008 ;
Schreuder et al. 2010 ).
BCI applications must also match their users’ preferences.
Carefully tailoring the application to the individual while
working within the constraints of the system design will
improve use acceptance and general satisfaction with the BCI.
Since motivation is critical in ensuring subject participation,
the choice of application is extremely important. For example,
people with high spinal-cord injuries who are still able to
speak, may not be interested in a BCI application that controls
a speech-generating device, but may be very interested in an
application that controls a computer mouse.
 e BCI applications that have been tested thus far in home
use are based mainly on selection of icons presented on a com-
puter screen, and they o en include sequential menu formats.
 ey can provide a number of simple functions, including
word processing, e-mail, environmental control, and Internet
access (e.g., Sellers et al. 2010 ). Menu formats and sequences
can be con gured to match the capacities, needs, and prefer-
ences of each user.  ey can support important functions such
as: requests for medical or other care; room temperature and
other environmental controls; answering simple questions (in
print or with a speech synthesizer); interactions with family
members or friends; requests for food or drink; e-mail; word-
processing; entertainment; Internet access; and others. Figure
20–3 shows an e-mail application that several users of the
Wadsworth P300-based BCI home system are now employing
to communicate with family and friends.
As discussed in detail in chapter 11, guidelines, standards,
and examples abound in the  eld of AT, and BCI researchers
20-Wolpaw-20.indd 329
20-Wolpaw-20.indd 329
8/10/2011 5:14:41 PM
8/10/2011 5:14:41 PM
OUP UNCORRECTED PROOF – FIRST-PROOF, 11/08/2011, GLYPH

330 | BRAIN–COMPUTER INTERFACES
85
84
83
82
81
80
79
78
77
76
75
74
73
72
71
70
69
68
67
66
65
64
63
62
61
60
59
58
57
56
55
54
53
52
51
50
49
48
47
46
45
44
43
42
41
40
39
38
37
36
35
34
33
32
31
30
29
28
27
26
25
24
23
22
21
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
should avail themselves of the extensive technology, experi-
ence, and expertise available in that  eld. Indeed, BCI home
systems are best viewed as technology that extends the spec-
trum of conventional (i.e., muscle-based) AT technology, and
BCIs will o en be most e ective when used as new control
interfaces for existing AT devices (chapter 11). BCI clinical
research can bene t from innovations in AT and in other areas
of human-computer interface (HCI) research and develop-
ment (e.g., Cook and Hussey 2002 ; Cremers et al. 1999 ).  ese
can be as straightforward as language-prediction programs
(Ryan et al. 2011 ) or as novel as the Hex-o-spell (Blankertz
et al. 2007 ; Williamson et al. 2009 ).
PLACING THE BCI IN THE HOME
In the transition from the laboratory to the home, many new
factors that can interfere with BCI use come into play (Sellers
et al. 2003 ; Sellers and Donchin 2006 ; Neumann and Kubler
2003 ). Although the nature of their vulnerabilities varies
with their methodology, all BCIs systems are likely to encoun-
ter a variety of di culties in making the transition from the
simple, highly controlled laboratory environment to much
more variable, uncontrolled, and demanding home environ-
ments.  is is likely to be the case for both noninvasive and
invasive BCIs and for BCIs that use electrical or metabolic sig-
nals. Because most BCI types remain largely con ned to the
laboratory, the discussion here necessarily focuses on the prob-
lems encountered by the EEG-based BCIs now being tested in
home use.
Figure 20–2 A shows a person with ALS using a P300-based
BCI. It is clear from the  gure that, in addition to the BCI
equipment, several other electronic and medical devices
including a ventilator are in very close proximity.  e clutter
typical of the immediate home environment of severely dis-
abled people (who are usually in a wheelchair or a bed with
various medical equipment close by) requires that the BCI
system be portable and su ciently small to  t into this com-
plex environment.  e typical home also has other distractions
(e.g., people entering and exiting the room, telephones ringing,
dogs barking, etc.) that may interfere with the attention needed
for BCI usage and that should also be considered in deciding
where to place the BCI. Working together, the user, caregiver,
and investigators should consider the setting(s) in which the
BCI will be used, and decide how the user and the system com-
ponents will best be situated.
 e typical home has multiple sources of electromagnetic
noise that can degrade the quality of EEG recording. In addi-
tion to generating ongoing 60-Hz (or 50-Hz) line noise, heat-
ing/cooling appliances (e.g., refrigerators) that cycle on and o
and other appliances such as electric garage-door openers can
produce severe transient artifacts.  e ventilators essential to
the survival of many prospective BCI users o en cause high-
frequency electromagnetic artifacts as well as low-frequency
mechanical (i.e., movement) artifacts (Young and Campbell
1999 ). Such electromagnetic noise can be reduced by proper
grounding and secure connection of the ground and reference
electrodes and by such maneuvers as suspending the electrode
cables or simply moving them away from the ventilator. Low-
frequency mechanical artifacts caused by head movement with
respiration may be reduced by simple solutions such as putting
additional padding or pillows behind the user’s head or dis-
pensing with the sponge pads sometimes placed under EEG
electrodes. Caregivers and others should be instructed to take
care not to disturb system components or cables once they are
properly placed. Finally, it may be necessary to eliminate
remaining artifacts (e.g., 60-Hz line noise) with  ltering meth-
ods (see chapter 7). Furthermore, in addition to addressing
sources of artifacts, it is important to ensure that the electrical
power in the home is su ciently stable. In some situations, use
of an uninterruptible power supply (UPS) may be necessary.
As each home environment is di erent, the various sources
of interference must be addressed on a case-by-case basis
(Sellers and Donchin 2006 ). To be suitable for home use, a BCI
system must be robust enough to avoid or accommodate these
problems. Determination of the extent to which a given system
meets this requirement is one of the key goals of a home
study.
Another important part of situating the BCI in the home is
resolving how the daily data on system operation and other
important data (e.g., periodic copy-spelling sessions for adjust-
ing system parameters and/or measuring accuracy) can be
transferred to the investigators remotely. Ideally, this can be
accomplished in an automated fashion through an internet
link. For example, this transfer may use remote desktop con-
trol (Cohen 2004 ). GoToMyPC
®
(Citrix Systems) is a service
that provides secure access to remote sites and was used for
transferring BCI data by Sellers et al. ( 2010 ). It supports data
Figure 20.3 The e-mail application for the P300-based Wadsworth BCI home
system. (A) On the right is the standard 8 × 9 matrix capable of controlling any
Windows-based program that can be operated with a keyboard. On the left is
an e-mail that the BCI user has just composed and sent. The green “Send”
confi rms to the user that the “Send” command has been recognized and
executed. The small window below the message is an optional predictive
speller feature that can increase writing speed. (B) On the left is the Help
Menu, which can be accessed by selecting the word “Help” from the bottom
row-fourth column of the matrix (right). This menu lists commands that can be
executed through other matrix selections. (See chapter 12 in this volume for
full explication of the P300-based BCI methodology used here.)
20-Wolpaw-20.indd 330
20-Wolpaw-20.indd 330
8/10/2011 5:14:41 PM
8/10/2011 5:14:41 PM
OUP UNCORRECTED PROOF – FIRST-PROOF, 11/08/2011, GLYPH

CHAPTER 20. CLINICAL EVALUATION OF BCIS | 331
113
112
111
110
109
108
107
106
105
104
103
102
101
100
99
98
97
96
95
94
93
92
91
90
89
88
87
86
85
84
83
82
81
80
79
78
77
76
75
74
73
72
71
70
69
68
67
66
65
64
63
62
61
60
59
58
57
56
55
54
53
52
51
50
49
48
47
46
45
44
43
42
41
40
39
38
37
36
35
34
33
32
31
30
29
28
27
26
25
24
23
22
21
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
transfer as well as real-time interaction. A separate license is
required for each site.
ENSURING SAFETY AND COMFORT
User safety and comfort and caregiver convenience are
extremely important and require close and comprehensive
attention. Many years of research and use in intensive care
units, operating rooms, and emergency rooms show that long-
term EEG use is compatible with ventilator technology
(Friedman et al. 2009 ; Phillips et al. 2010 ; Tantum 2001 ). BCI
clinical researchers must ensure that BCI presence and use
does not a ect the functioning of other important medical
devices. Prior to home installation, each BCI home system, like
all medical equipment, should undergo a formal safety evalua-
tion by a hospital electronics support group or similar body.
Furthermore, users and caregivers need to understand that
BCIs do not substitute for standard monitoring of the BCI user
who has compromised pulmonary functions and thus that
ventilator alarms and other safeguards must remain in place
(Fludger and Klein 2008 ). In designing a BCI system and its
clinical study, it is also important to eliminate to the greatest
extent possible the chance that BCI (or user) malfunctions
might compromise safety (see chapter 24). For example, stud-
ies that enable independent use of environmental controls
should ensure that the BCI cannot produce outputs that could
endanger the user (e.g., by setting the room temperature too
high). All the tasks that the BCI enables should be structured
to prevent their creation of safety hazards.
For EEG-based studies, there is an extremely small chance
of skin abrasion.  is risk depends on the particular sensor cap
and gel.  e Wadsworth Center BCI research group has used
the Electro-Cap International
TM
cap system for 5000 + hours in
the lab, has monitored 1000 + hours of its independent home
use, and has not encountered a single incident of such abra-
sion. Despite this reassuring experience, researchers and care-
givers must remain alert to the possibility, and caregivers
should make regular scalp inspection part of their normal BCI
r o u t i n e .
TRAINING THE USER AND THE CAREGIVERS
In the course of the initial BCI evaluations and demonstrations
of the available applications, the user typically becomes famil-
iar with the basic features of BCI use. Nevertheless, to ensure
that di culties do not arise from simple misunderstandings or
inadequate orientation, researchers should provide guided
practice and well-documented help menus.  e more chal-
lenging and complex requirement is training the caregiver to
support BCI system use. It is essential to have a logical and
complete caregiver training protocol. Caregivers must know
how to initiate and oversee e ective BCI operation. Since fully
asynchronous BCIs are not yet available for home use (see
chapter 10), the initiation of BCI usage requires substantial
neuromuscular function, and thus it involves a caregiver.
 e caregiver must learn how to: place the electrode cap on
the user so that it is comfortable and properly positioned; add
electrode gel; turn the BCI system on; check that all electrodes
are recording good EEG signals and  x any that are not; initiate
system use; monitor BCI operation; turn the system o ; remove
the cap and maintain the cap and electrodes in good working
order; recognize technical problems or poor performance and
request technical support as needed; ensure that data transfer
to the research lab occurs as required; and ensure that periodic
brief copy-spelling sessions for checking system parameters
and/or measuring performance take place.
Typically, the caregiver’s training will occupy two or three
separate 1-hour sessions and will culminate with the investiga-
tor simply watching the caregiver go through the entire BCI
usage process (i.e., placing the cap and starting the system,
overseeing operation, removing and cleaning the cap), as well
as the ancillary processes (e.g., data transfer, copy-spelling
session).
Neither caregivers, users, nor other clinical personnel are
likely to be trained researchers.  erefore, all information,
even for routine tasks, should be carefully scripted. Each train-
ing objective (e.g., cap placement, skin preparation, gel appli-
cation, electrode check, etc.) should be demonstrated and then
practiced, with training objectives clearly described and pro -
ciency for each task tested separately (Gursky and Ryser 2007 ).
For the caregiver, the required objectives may include some
that are seemingly obvious but nonetheless crucial (e.g., con-
tinuing to devote his or her attention to the user while follow-
ing the instructions on the screen). In addition to initiating
and stopping the BCI, the caregiver should also be able to
pause and resume BCI operation for essential activities (e.g.,
tracheal suction for a user who is on a ventilator [C. Wolf, per-
sonal communication, 2011]).
Figure 20–2 B displays a tool used to train the caregiver and
to serve as a reminder that electrode impedances must be
below an acceptable level prior to starting the BCI system.
Eight circles representing the electrodes can be red, yellow, or
green. Green indicates acceptable impedance. Yellow or red
indicates that the electrode needs further attention (e.g., skin
preparation, gel). Other screens provide guidance in placing
the cap, and testing the connections between the computer and
the ampli er and monitor. As a general rule, caregiver training
is most likely to be successful when the complexity of the
hardware and so ware are minimized to the greatest extent
p o s s i b l e .
PROVIDING ONGOING TECHNICAL SUPPORT
AS NEEDED
Once the BCI is placed in the home, and the user and
caregiver(s) are adequately trained, independent daily use can
begin.  roughout this use, and particularly in the initial weeks
and months, the investigators should closely monitor opera-
tion remotely and be readily available to resolve any di culties
that arise.  is oversight is essential for gathering the basic
data of the study and also for maximizing the likelihood that
the BCI will come to serve important purposes in the user’s
daily life.  e system will be used only if it works reliably and
with minimal di culty.  us, it is crucial, particularly in the
early days, for the investigators to respond quickly to any
problems that arise, and to be prepared to correct them
immediately.
Many problems may be resolved remotely, through e-mail
or phone discussions with the caregiver, analyses of data sent
20-Wolpaw-20.indd 331
20-Wolpaw-20.indd 331
8/10/2011 5:14:42 PM
8/10/2011 5:14:42 PM
OUP UNCORRECTED PROOF – FIRST-PROOF, 11/08/2011, GLYPH

332 | BRAIN–COMPUTER INTERFACES
113
112
111
110
109
108
107
106
105
104
103
102
101
100
99
98
97
96
95
94
93
92
91
90
89
88
87
86
85
84
83
82
81
80
79
78
77
76
75
74
73
72
71
70
69
68
67
66
65
64
63
62
61
60
59
58
57
56
55
54
53
52
51
50
49
48
47
46
45
44
43
42
41
40
39
38
37
36
35
34
33
32
31
30
29
28
27
26
25
24
23
22
21
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
over the Internet, or real-time audiovisual interactions over
the Internet. Others may require home visits, and (rarely)
replacement of a system component. It is worthwhile, and
might be considered a key aspect of a BCI home-use study, to
employ a formal system for documenting problems and the
time and e ort involved in their solution. Such data are impor-
tant in assessing the clinical (and ultimately the  nancial)
practicality of the BCI system.
To a signi cant degree, problems may be reduced by care-
ful selection of system components and prophylactic measures
aimed at ensuring that they function satisfactorily as long as
possible. For example, one of the most widely used EEG caps
(ElectroCap, Inc.) has been estimated to have an average life
span of 450 hours.  is corresponds to 450 diagnostic sessions
in a clinical EEG laboratory. However, a home BCI system
might be used 5 hr/day, 7 days/week, which is 1820 hr/year
(Sellers et al. 2010 ).  us, several caps are likely to be needed
by an individual home user each year. Careful cleaning and
regular cap rotation may extend cap and electrode life span
and reduce the incidence of poor BCI performance caused by
cap or electrode malfunction. Nevertheless, for a person who
uses the BCI many hours per day, caps should be routinely
replaced or refurbished every few months, rather than simply
changed when they fail.
As time passes, and the skills and sophistication of the user
and caregiver increase, problems are likely to arise less fre-
quently. Nevertheless, it is prudent to continue periodic regu-
lar home visits, even if at relatively long intervals. During such
visits, the user’s physical state and environment may be reas-
sessed, applications may be added or upgraded as appropriate,
and adjustments may be made in the BCI hardware and
con guration.
MEASURING THE EXTENT, NATURE, AND
SUCCESS OF BCI USE
 e automated transfer of complete data on BCI system opera-
tion should allow full quanti cation of the extent (i.e., days
used, hours/day) and nature (i.e., speci c applications) of
daily BCI use.  e measurement of performance, speci cally
accuracy, is more problematic because, for most routine usage,
the actual intention of the user (i.e., the correct BCI output)
is not known with certainty. Periodic brief copy-spelling ses-
sions in which the system speci es the correct output are the
most straightforward solution. Alternatively, or in addition,
appropriate analysis programs (designed with appropriate
attention to user privacy concerns (see chapter 24) may detect
errors (e.g., spelling mistakes in written text) and calculate
accuracies.
It is also important to monitor other aspects of the user’s
state and environment for changes that may greatly a ect BCI
use. Disturbances such as intercurrent illnesses may interrupt
the user’s normal routine and can greatly reduce BCI use, at
least temporarily. Other problems, such as the temporary
absence or permanent departure of the caregiver who supports
BCI use, and the need to train a replacement, may also reduce
BCI use. Fluctuations or progression in the user’s basic disease,
particularly for users with ALS, may also a ect BCI use. For
people with ALS, monitoring of this progression may be
accomplished with the revised ALS functional rating scale
(ALSFRS R) which provides a succinct measure of disability
(Cedarbaum et al. 1999 ). In addition to standard monitoring
of these speci c factors that may a ect BCI use, caregivers and
investigators should be alert to sudden changes in BCI use that
might be caused by changes in the user’s physical or mental
state or in other factors.
Finally, periodic questionnaire-based interviews of users,
caregivers, and family members are useful ancillary tools for
identifying system or procedure modi cations that might
improve BCI performance or usefulness and/or increase user
or caregiver satisfaction and convenience.
DOES THE BCI IMPROVE THE USER’S LIFE?
Certainly, the simplest and most obvious measure of BCI use-
fulness is the extent to which it is used. No matter how simple
and convenient, BCI use requires signi cant commitment on
the part of both the user and the caregiver.  us, frequent use
is probably a good indicator that the user  nds it worthwhile.
At the same time, for scienti c evaluation, the validation of a
home BCI system requires more formal and substantive assess-
ment of its impact on the lives of its users and their caregivers,
as well as on their family and friends.
Recent studies indicate that, despite common assumptions,
quality of life (QoL) can be quite good in people with severe
motor disabilities (e.g., Kubler et al. 2005b ; Nygren and
Askmark 2006 ; Chio et al. 2004 ; Simmons et al. 2006 ). Indeed,
this  nding provides much of the impetus for BCI develop-
ment.  e measures developed for these QoL studies can also
be used to evaluate the impact of BCI use.
One of the most important considerations in choosing an
assessment instrument is its length. To ensure accurate and
complete data collection from individuals who may have di -
culty communicating, any instrument should be relatively
brief. One such instrument is the McGill QoL questionnaire,
which was designed for individuals with advanced disease
(Cohen et al. 1995 ; Cohen et al. 1996 ). It is widely used as the
basis for other, more elaborate questionnaires, including the
Simmons scale designed speci cally for ALS (Simmons et al.
2006 ).  e McGill questionnaire consists of 17 questions in
two parts. Part A consists of one comprehensive question
asking the patient for an overall assessment of his/her quality
of life (and is itself capable of providing a basic QoL measure).
Part B includes 16 questions that cover physical, psychological,
existential, and support domains. Answers are indicated on an
11-point Likert scale (0–10). Depending on practicality, addi-
tional more complex measures might be used to assess QoL in
BCI users with severe disabilities (e.g., Chio et al. 2004 ; Kubler
et al. 2007; Kurt 2007 ; Lulé et al., 2009 ; Mautz et al. 2010 ;
Simmons et al. 2006 ).
In addition to the impact on the BCI users, comparable
measures are available for evaluating BCI impact on others
(e.g., caregivers, family members), as well for evaluating others’
perceptions of how the BCI is a ecting the user.  ese instru-
ments (e.g.,  e Psychosocial Impact of Assistive Devices Scale
[PIADS] [Derosier and Farber 2005 ; Giesbrecht et al. 2009 ;
Scherer et al. 2010]) may be administered at the beginning of
20-Wolpaw-20.indd 332
20-Wolpaw-20.indd 332
8/10/2011 5:14:42 PM
8/10/2011 5:14:42 PM
OUP UNCORRECTED PROOF – FIRST-PROOF, 11/08/2011, GLYPH

CHAPTER 20. CLINICAL EVALUATION OF BCIS | 333
113
112
111
110
109
108
107
106
105
104
103
102
101
100
99
98
97
96
95
94
93
92
91
90
89
88
87
86
85
84
83
82
81
80
79
78
77
76
75
74
73
72
71
70
69
68
67
66
65
64
63
62
61
60
59
58
57
56
55
54
53
52
51
50
49
48
47
46
45
44
43
42
41
40
39
38
37
36
35
34
33
32
31
30
29
28
27
26
25
24
23
22
21
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
the study and at intervals of some months therea er. Positive
changes in these measures can constitute important evidence
for the practical clinical value of a BCI system.
BCI e cacy may also be measured in other ways, such as
by its ability to permit reductions in caregiver e ort, or to
increase the productivity of the user. For example, the inde-
pendent communication enabled by present-day P300-based
BCIs may free the caregiver from serving as a communication
partner (and the user from the need to have a partner), or may
even help the user to continue productive employment (e.g.,
Sellers et al. 2010 ).
DIFFICULT CHALLENGES IN BCI
TRANSLATIONAL STUDIES
BCI translational studies confront  ve di cult challenges that
arise from the nature of the user population. First, the users are
typically extremely disabled and may have progressive diseases.
 eir highly compromised physical states, medication regi-
mens, frequent intercurrent illnesses, and dependence on o en
transient caregivers mean that many factors unrelated to the
BCI system itself may greatly a ect its day-to-day usage and
distort the data that quantify that use. Furthermore, for people
with progressive disease (e.g., ALS, multiple sclerosis), their
overall level of function and their need for and ability to use
the BCI may change markedly over the course of the study.
 is may further complicate the task of assessing BCI impact.
In the case of ALS particularly, a substantial number of users
may die in the course of a long-term study (Murray 2006 ).
 e second issue is that it is extremely di cult or even
wholly impractical to conduct large-scale fully controlled stud-
ies that compare BCIs to conventional assistive technology.
 e number of appropriate subjects is limited and the partici-
pation of each one requires prolonged e ort on the part of the
investigators.  us, studies implemented by a single laboratory
will generally have small numbers of subjects. Although coor-
dinated multicentered studies are a possible method for study-
ing many subjects, they require an expensive and demanding
second level of organization and oversight to ensure unifor-
mity of subject selection, investigator training, and study exe-
cution across the multiple sites involved. Furthermore,
controlled studies comparing BCI systems with other assistive
technology (e.g., eye-gaze systems) introduce further complex-
ity in terms of standardization of methods and uniformity of
procedures. One potential response to this problem is a study
design in which each subject serves as his or her own control
(i.e., uses the BCI for 6 months, then an eye-gaze system or
nothing for 6 months, etc.) However, such designs are likely to
be di cult to justify ethically (much less implement) in
extremely disabled users, and they may be essentially impos-
sible in users with progressive disorders such as ALS.
 e third issue concerns study duration and long-term
commitment to the user. In general, formal studies usually
specify a time period over which each subject is studied (e.g., 1
or 2 years in the case of home BCI use). However, if the BCI is
successful, that is, if it substantially improves the user’s life,
s/he may very understandably want to continue to use it.
Indeed, this is particularly probable for the extremely disabled
subjects who are the users of BCI systems now ready for clini-
cal testing. Since these BCI systems are relatively inexpensive,
simply allowing the user to keep the hardware past the end of
the study may not be a major problem. However, the continu-
ing need for technical support and supplies (e.g., electrode
caps) requires continued funding as well as expertise that may
be available only from the laboratory that conducted the study,
which means that the laboratory personnel need to be available
and able to provide the support. Although this problem will
presumably be resolved when BCI systems ultimately become
reimbursable medical devices, studies with nonreimbursable
systems are currently needed to provide the data that will jus-
tify such reimbursement.
 e issue of commitment is even more complex for sub-
jects who have progressive disorders.  e BCI may serve them
well initially, but as their disease progresses it may become
ine ective.  e subject, caregivers, or family members may
then ask or expect the investigators to modify the system so
that it can continue to function e ectively. Although the inves-
tigators may indeed want to do this, they may lack the requisite
resources or expertise. At this point no general solution
is apparent for these very di cult situations, and acceptable
courses of action must be developed on a case-by-case basis. It
behooves the investigators to anticipate these situations as they
design BCI studies and to consider how they might respond
most e ectively (see chapter 24 for further discussion).
 e fourth issue concerns subjects who may well need a
BCI, but who do not qualify for the study or cannot use the
BCI system under study.  e ad hoc development of new mod-
i cations to accommodate a single prospective user (beyond
those possible in the existing system or readily implemented,
such as covering one eye to prevent diplopia) is likely to divert
investigator e orts and resources from the study itself and
unlikely to be successful. Furthermore, such modi cations
may well constitute entirely new research endeavors that
require their own IRB reviews and approvals. In general, if a
clinical study is to be carried forward to completion and to
yield substantive results, the range of possible subject-speci c
adjustments (e.g., matrix brightness, stimulus rate, etc.) should
be de ned from the start. Subjects who cannot achieve ade-
quate accuracy within this range of adjustments should not be
included in the study, painful though it may be to all involved
including the investigators. (At the same time, the investigators
might still o er substantive help to such individuals as
described in chapter 24.)
Finally, the need for initial and ongoing technical expertise
o en prevents the undertaking of BCI clinical studies alto-
gether, or limits them to individuals or institutions with sub-
stantial resources and very strong commitments to the
endeavor.  e development of e ective translational partner-
ships like that undertaken between the BCI research group at
the Wadsworth Center and clinicians at the Helen Hayes
Rehabilitation Hospital can enable BCI clinical studies. Such
partnerships between researchers and clinicians may facilitate
and accelerate the translation of BCI systems from the labora-
tory to successful long-term home use by those who need
them.
20-Wolpaw-20.indd 333
20-Wolpaw-20.indd 333
8/10/2011 5:14:42 PM
8/10/2011 5:14:42 PM
OUP UNCORRECTED PROOF – FIRST-PROOF, 11/08/2011, GLYPH

334 | BRAIN–COMPUTER INTERFACES
119
118
117
116
115
114
113
112
111
110
109
108
107
106
105
104
103
102
101
100
99
98
97
96
95
94
93
92
91
90
89
88
87
86
85
84
83
82
81
80
79
78
77
76
75
74
73
72
71
70
69
68
67
66
65
64
63
62
61
60
59
58
57
56
55
54
53
52
51
50
49
48
47
46
45
44
43
42
41
40
39
38
37
36
35
34
33
32
31
30
29
28
27
26
25
24
23
22
21
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
FUTURE IMPROVEMENTS THAT WILL
PROMOTE BCI CLINICAL TRANSLATION
 e practicality and appeal of EEG-based BCI systems for
home use should be greatly augmented by the continuing
development both of more streamlined hardware and so ware
and of applications that are useful to people with severe dis-
abilities (e.g., Cincotti et al. 2008 ; Münßinger et al. 2010 ; Sellers
et al. 2010 ). In addition, convenience and comfort can be
increased by the development of dry or active electrode sys-
tems (e.g., Gargiulo et al. 2010 ; Popescu et al. 2007 ; Sellers et al.
2009 ; see also chapter 6). Cosmesis can be improved with more
attractive and/or inconspicuous electrode mountings (i.e.,
electrode caps that look like ordinary hats or helmets).
Although the standard electrode cap with gel application func-
tions adequately, gel-free electrodes and more comfortable
caps are clearly important to many prospective users. Smaller
more robust ampli ers and computers and replacement of
wired connections with telemetry should further increase the
convenience, cosmesis, portability, and durability of these
systems. Decreases in the complexity of the system hardware
(e.g., number of electrodes) and so ware, and increase in
reliability, speed, and range of useful applications will also
encourage BCI home use.
SUMMARY
BCIs are fast becoming e ective communication and control
devices. However, they are still con ned almost entirely to the
protected environments of a multitude of laboratories through-
out the world.  is focus leaves a major research gap that must
be addressed if BCIs are to ful ll their primary purpose and jus-
tify the considerable support their development receives from
governments and other funding entities.  e BCIs that work well
in the laboratory need to be shown to work well in real life and to
provide to people with disabilities new communication and
capabilities that improve their daily lives. To meet these require-
ments, they must be simple to operate, need minimal expert
oversight, be usable by people who are extremely disabled, and
provide reliable, long-term performance in complex home envi-
ronments.  eir capacity to satisfy these demanding criteria can
only be determined through studies of their long-term perfor-
mance in independent daily home use by the people with severe
disabilities who constitute their target user population.
Once a BCI has proven itself in the laboratory, the transla-
tional research that seeks to establish its clinical usefulness
must address four questions:
Can the BCI be implemented in a form suitable
•
for long-term home use?
Who needs and can use the BCI?
•
Can her/his home environment support the •
BCI usage and does she/he actually use it?
Does the BCI improve his/her life?
•
 is chapter reviews the multiple complex issues involved
in addressing each of these questions.  ese include: BCI
system robustness, convenience, and portability; subject inclu-
sion criteria; informed consent; the suitability of the home
environment; user and caregiver education and training; user-
speci c system con guration and applications; ongoing tech-
nical support; collection of data on amount, type, and success
of BCI usage; complications by intercurrent illness and care-
giver changes; and evaluation of impact on user quality of life.
 e chapter also addresses di cult issues particularly relevant
to BCI studies, including disease progression, the practical
limitations on controls and on the size of study populations,
and the issues that may arise when time-limited studies end.
REFERENCES
Albert , S. M. , A. Whitaker , J. G. Rabkin , M. del Bene , T. Tider , I. O’Sullivan ,
and H. Mitsumoto . 2009 . Medical and supportive care among people with
ALS in the months before death or tracheostomy . J Pain Symptom Manage
38 ( 4 ): 546 – 553 .
A l l e n , K . D . , E . J . K a s a r s k i s e , R . S . B e d l a c k , M . P. R o z e a r , J . C . M o r g e n l a n d e r , A .
S a b e t g , L . S a m s , J . H . L i n d q u i s t , M . L . H a r r e l s o n a , C . J . C o  man, and E. Z.
Oddone . 2008 ,  e National Registry of Veterans with Amyotrophic
Lateral Sclerosis . Neuroepidemiology 30 : 180 – 190 .
Bai , O., P. Lin , D. Huang , D. Y. Fei and M. K. Floeter . 2010 . Towards a
user-friendly brain-computer interface: initial tests in ALS and PLS
patients . Clin Neurophysiol 121 ( 8 ): 1293 – 1303 .
Bauernfeind , G. , R. Leeb , S. C. Wriessnegger , and G. Pfurtscheller . 2008 .
Development, set-up and  rst results for a one-channel near-infrared
spectroscopy system . Biomed Tech (Berl) 53 ( 1 ): 36 – 43 .
Bedlack , R.S. , P. Wicks , J. Heywood , and E. Kasarskis . 2010 . Modi able barri-
ers to enrollment in American ALS research studies . Amyotroph Lat Scler
11 ( 6 ): 502 – 507 .
B i r b a u m e r , N . , N . G h a n a y i m , T . H i n t e r b e r g e r , I . I v e r s e n , B . K o t c h o u b e y , A .
Kubler , J. Perelmouter , E. Taub , and H. Flor . 1999 . A spelling device for the
paralysed . Nature 398 ( 6725 ): 297 – 298 .
Black , B. B. , P. V. Rabins , J. Sugarman , and J. H. Karlawish , 2010 . Seeking assent
and respecting dissent in dementia research. Am J Geriatr Psychiatry
18 ( 1 ): 77 – 85 .
B l a n k e r t z , B . , M . K r a u l e d a t , G . D o r n h e g e , J . W i l l i a m s o n , R . M u r r a y - S m i t h , a n d
K.-R. Müller . 2007 , A note on brain actuated spelling with the Berlin brain-
computer interface . In C. Stephanidis (Ed.), Universal Access in Human-Com-
puter Interaction. Ambient Interaction , Berlin : Springer-Verlag , 4555 : 759 – 768 .
Bradshaw , L. A. , R. S. Wijesinghe and J. P. Wikswo , Jr . 2001 . Spatial  lter
approach for comparison of the forward and inverse problems of elec-
troencephalography and magnetoencephalography . Ann Biomed Eng
29 ( 3 ): 214 – 226 .
B u c h , E . , C . W e b e r , L . G . C o h e n , C . B r a u n , M . A . D i m y a n , T . A r d , J . M e l l i n g e r ,
A. Caria , S. Soekadar , A. Fourkas , and N. Birbaumer . 2008 .  ink to move:
a neuromagnetic brain-computer interface (BCI) system for chronic
s t r o k e . Stroke 39 ( 3 ): 910 – 917 .
C e d a r b a u m , J . M . , N . S t a m b l e r , E . M a l t a , C . F u l l e r , D . H i l t , B .  urmond , and
A. Nakanishi . 1999 .  e ALSFRS-R: a revised ALS functional rating scale
that incorporates assessments of respiratory function . BDNF ALS Study
Group (Phase III). J Neurol Sci 169 ( 1–2 ): 13 – 21 .
C h i o , A . , A . G a u t h i e r , A . M o n t u s c h i , A . C a l v o , N . D i V i t o , P . G h i g l i o n e , a n d R .
Mutani . 2004 . A cross sectional study on determinants of quality of life in
ALS . J Neurol Neurosurg Psychiatry 75 ( 11 ): 1597 – 1601 .
C i n c o t t i , F . , D . M a t t i a , F . A l o i s e , S . B u f a l a r i , G . S c h a l k , G . O r i o l o , A . C h e r u b i n i ,
M. G. Marciani , and F. Babiloni . 2008 . Non-invasive brain-computer inter-
face system: towards its application as assistive technology . Brain Res Bull
75 ( 6 ): 796 – 803 .
Cohen , D . 1972 . Magnetoencephalography: detection of the brain’s electrical
activity with a superconducting magnetometer . Science 175 ( 22 ): 664 – 666 .
Cohen , S. R. , B. M. Mount , M.G. Strobel , and F. Bui . 1995 .  e McGill Quality
of Life Questionnaire: a measure of quality of life appropriate for people
with advanced disease. A preliminary study of validity and acceptability .
Palliative Med 9 ( 3 ): 207 – 219 .
Cohen , S R. , B. M Mount , J. J. Tomas , and L. F. Mount . 1996 . Existential well-
being is an important determinant of quality of life. Evidence from the
McGill Quality of Life Questionnaire . Cancer
. 77 ( 3 ): 576 – 586 .
20-Wolpaw-20.indd 334
20-Wolpaw-20.indd 334
8/10/2011 5:14:42 PM
8/10/2011 5:14:42 PM
OUP UNCORRECTED PROOF – FIRST-PROOF, 11/08/2011, GLYPH

CHAPTER 20. CLINICAL EVALUATION OF BCIS | 335
152
151
150
149
148
147
146
145
144
143
142
141
140
139
138
137
136
135
134
133
132
131
130
129
128
127
126
125
124
123
122
121
120
119
118
117
116
115
114
113
112
111
110
109
108
107
106
105
104
103
102
101
100
99
98
97
96
95
94
93
92
91
90
89
88
87
86
85
84
83
82
81
80
79
78
77
76
75
74
73
72
71
70
69
68
67
66
65
64
63
62
61
60
59
58
57
56
55
54
53
52
51
50
49
48
47
46
45
44
43
42
41
40
39
38
37
36
35
34
33
32
31
30
29
28
27
26
25
24
23
22
21
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
Cohen , T . 2004 . Medical and information technologies converge . IEEE Eng
Med Biol Mag 23 ( 3 ): 59 – 65 .
Conradi , J. , B. Blankertz , M. Tangermann , V. Kunzmann , and Curio G . 2009 .
Brain-computer interfacing in tetraplegic patients with high spinal cord
injury . Int J Bioelectromag 11 ( 2 ): 65 – 68 .
Cook , A. M. , and S. M. Hussey . 2002 . Assistive Technologies: Principles and
Practice . St. Louis : Mosby .
Cotterell , P . 2008 . Striving for independence: experiences and needs of service
users with life limiting conditions . J Adv Nurs 62 ( 6 ): 665 – 673 .
Coyle , S. M. , T. E. Ward , and C. M. Markham . 2007 . Brain-computer interface
using a simpli ed functional near-infrared spectroscopy system . J Neural
Eng 4 ( 3 ): 219 – 226 .
Cremers , G. , H. Kwee , and M. Sodoe . 1999 . Interface design for severely
disabled people . In C. Buehler and H. Knops (Eds.), Assistive Technology
Research Series . Washington, DC : IOS Press .
Derosier , R. and R. S. Farber . 2005 . Speech recognition so ware as an assistive
device: a pilot study of user satisfaction and psychosocial impact . Work
25 ( 2 ): 125 – 134 .
Donoghue , J. P . 2008 . Bridging the brain to the world: a perspective on neural
interface systems . Neuron 60 : 511 – 521 .
F a r q u h a r , J . , J . B l a n k e s p o o r , R . V l e k , a n d P. D e s a i n . T o w a r d s a n o i s e - t a g g i n g
auditory BCI-paradigm . Proceedings of the 4th International Brain-
Computer Interface Workshop and Training Course . 2008 : 50 – 55 .
Farwell , L. A. , and E. Donchin . 1988 . Talking o the top of your head: toward
a mental prosthesis utilizing event-related brain potentials . Electroen-
cephalogr clin Neurophysiol 70 ( 6 ): 510 – 523 .
Fludger , S. , and A. Klein . 2008 . Portable ventilators . Anaesth, Critl Care Pain
8 ( 6 ): 199 – 203.
Friedman , D. , J. Claassen , and L.J. Hirsh . 2009 . Continuous electroencephalo-
gram monitoring in the intensive care unit . Anesth Analg. 109 ( 2 ):
506 – 523 .
Furdea , A. , S. Halder , D. J. Krusienski , D. Bross , F. Nijboer , N. Birbaumer , and
A. Kubler . 2009 . An auditory oddball (P300) spelling system for brain-
computer interfaces . Psychophysiology 46 ( 3 ): 617 – 625 .
G a r g i u l o , G . , R . A . C a l v o , P. B i f u l c o , M . C e s a r e l l i , C . J i n , A . M o h a m e d a n d A .
van Schaik . 2010 . A new EEG recording system for passive dry electrodes .
Clin Neurophysiol , 121 : 686 – 693 .
G a u t h i e r , A . , A . V i g n o l a , A . C a l v o , E . C a v a l l o , C . M o g l i a , L . S e l l i t t i , R . M u t a n i ,
and A. Chio . 2007 . A longitudinal study on quality of life and depression in
ALS patient-caregiver couples . Neurology 68 ( 12 ): 923 – 926 .
Giesbrecht , E.M. , J. D. Ripat , A. O. Quanbury and J. E. Cooper . 2009 ,
Participation in community-based activities of daily living: comparison of
a pushrim-activated, power-assisted wheelchair and a power wheelchair .
Disabil Rehabil Assist Technol 4 ( 3 ): 198 – 207 .
Guo , J. , S. Gao , and B. Hong . 2010 . An auditory brain-computer interface using
active mental response . IEEE Trans Neural Syst Rehabil Eng 18 ( 3 ):
230 – 235 .
Gursky B.S. , and B. J. Ryser , 2007 . A training program for unlicensed assistive
personnel . J School Nurs 23 : 92 – 97 .
Hill , N. J. , T. N. Lal , K. Bierig , N. Birbaumer, and B. Schölkopf . 2005 . An audi-
tory paradigm for brain-computer interfaces . Adv Neural Info Proc Syst
17 : 569 – 576 .
H i n t e r b e r g e r , T . , N . N e u m a n n , M . P h a m , A . K u b l e r , A . G r e t h e r , N . H o f m ayer ,
B. Wilhelm , H. Flor , and N. Birbaumer . 2004 . A multimodal brain-based
feedback and communication system . Exp Brain Res 154 ( 4 ): 521 – 526 .
H o c h b e r g , L . R . , M . D . S e r r u y a , G . M . F r i e h s , J . A . M u k a n d , M . S a l e h , A . H.
Caplan , A. Branner , D. Chen , R. D. Penn , and J. P. Donoghue . 2006 .
Neuronal ensemble control of prosthetic devices by a human with tetraple-
gia . Nature 442 ( 7099 ): 164 – 171 .
H o  mann , U. , J. M. Vesin , T. Ebrahimi , and K. Diserens . 2008 . An e cient
P300-based brain-computer interface for disabled subjects . J Neurosci
Methods 167 ( 1 ): 115 – 125 .
Ikegami , S. , K. Takano , N. Saeki , and K. Kansaku . 2011 . Operation of a P300-
based brain-computer interface by individuals with cervical spinal cord
injury . Clin Neurophysiol 122 : 991 – 996 .
Kaiser , J. , F. Walker , S. Leiberg , and W. Lutzenberger . 2005 . Cortical oscillatory
activity during spatial echoic memory . Eur J Neurosci 21 ( 2 ): 587 – 590 .
Kanoh , S. , K. Miyamoto , and T. Yoshinobu . 2008 . A brain-computer interface
(BCI) system based on auditory stream segregation . Conf Proc IEEE Eng
Med Biol Soc 2008 : 642 – 645 .
K a u h a n e n , L . , P. J y l a n k i , J . L e h t o n e n , P. R a n t a n e n , H . A l a r a n t a , a n d M . S a m s .
2007 . EEG-based brain-computer interface for tetraplegics . Comput Intell
Neurosci Published online doi: 10.1155/2007/23864.
Kennedy , P. R. , and R. A. Bakay . 1998 . Restoration of neural output from a
paralyzed patient by a direct brain connection . Neuroreport 9 ( 8 ):
1707 – 1711 .
K i r c h h o  , K. T. , and K. A. Kehl . 2007 . Recruiting participants in end-of-life
r e s e a r c h . Am J Hosp Palliat Care 24 ( 6 ): 515 – 521 .
Kleih , S. C. , F. Nijboer , S. Halder , and A. Kubler . 2010 . Motivation modulates
the P300 amplitude during brain-computer interface use . Clin Neurophysiol
121 ( 7 ): 1023 – 1031 .
K l o b a s s a , D . S . , T . M . V a u g h a n , P. B r u n n e r , N . E . S c h w a r t z , J . R . W o l p a w , C.
Neuper , and E. W. Sellers . 2009 . Toward a high-throughput auditory P300-
based brain-computer interface . Clin Neurophysiol 120 ( 7 ): 1252 – 1261 .
Kubler , A. , A. Furdea , S. Halder , E. M. Hammer , F. Nijboer , and B. Kotchoubey .
2009 . A brain-computer interface controlled auditory event-related poten-
tial (p300) spelling system for locked-in patients . Ann N Y Acad Sci
1157 : 90 – 100 .
K u b l e r , A . , B . K o t c h o u b e y , J . K a i s e r , J . R . W o l p a w , a n d N . B i r b a u m e r . 2 0 0 1 .
Brain-computer communication: unlocking the locked in . Psychol Bull
127 ( 3 ): 358 – 375 .
K u b l e r , A . , F. N i j b o e r , J . M e l l i n g e r , T . M . V a u g h a n , H . P a w e l z i k , G . S c h a l k , D . J .
McFarland , N. Birbaumer , and J. R. Wolpaw . 2005a . Patients with ALS can
use sensorimotor rhythms to operate a brain-computer interface . Neurology
64 ( 10 ): 1775 – 1777 .
K u b l e r , A . , S . W i n t e r , A . C . L u d o l p h , M . H a u t z i n g e r , a n d N . B i r b a u m e r . 2 0 05b .
Severity of depressive symptoms and quality of life in patients with amyo-
trophic lateral sclerosis . Neurorehabil Neural Repair 19 ( 3 ): 182 – 193 .
Kurt , A. , F. Nijboer , T. Matuz , and A. Kubler . 2007 . Depression and anxiety in
individuals with amyotrophic lateral sclerosis: epidemiology and manage-
m e n t . CNS Drugs. 21 ( 4 ): 279 – 291 .
Lancet Neurology Editorial . 2009 . National registry o ers new hope for ALS .
Lancet Neurol . 8 : 1 .
Lee , J. H. , J. Ryu , F. A. Jolesz , Z. H. Cho , and S. S. Yoo . 2009 . Brain-machine
interface via real-time fMRI: preliminary study on thought-controlled
robotic arm . Neurosci Lett 450 ( 1 ): 1 – 6 .
L u l é , D . , C . Z i c k l e r , S . H ä c k e r , M . A . B r u n o , A . D e m e r t z i , F . P ellas , S. Laureys
and A. Kübler . 2009 . Life can be worth living in locked-in syndrome .
Prog
Brain Res 177 : 339 – 351 .
Matuz , T. , N. Birbaumer , M. Hautzinger , and A. Kubler . 2010 Coping with
amyotrophic lateral sclerosis: an integrative view . J Neurol Neurosurg
Psychiatry , 81 ( 8 ): 893 – 898 .
McCane , L.M. , T. M. Vaughan , D. J. McFarland , D. Zeitlin , L. Tenteromano ,
J . M a k , E . W. S e l l e r s , G . T o w n s e n d , C . S . C a r m a c k , a n d J . R . Wo l p a w .
2009 . Evaluation of individuals with ALS for in home use of a P300
brain computer . In Program No. 664.7/DD38. 2009 Neuroscience Meet-
ing Planner : Society for Neuroscience, Online ( http://www.sfn.org/index.
aspx?pagename=abstracts_am2009 ) .
McFarland , D. J. , W. A. Sarnacki , and J. R. Wolpaw . 2010 . Electroencephalographic
(EEG) control of three-dimensional movement . J Neural Eng 7 ( 3 ): 036007 .
M e l l i n g e r , J . , G . S c h a l k , C . B r a u n , H . P r e i s s l , W. R o s e n s t i e l , N . B i r baumer , and
A. Kubler . 2007 . An MEG-based brain-computer interface (BCI) .
Neuroimage 36 ( 3 ): 581 – 593 .
M i l l á n , J . d e l R . , R . R u p p , G . M ü l l e r - P u t z , R . M u r r a y - S m i t h , C . G i u g l i e m m a , M
T a n g e r m a n n , C . V i d a u r r e , F . C i n c o t t i , A . K ü b l e r , R . L e e b , C . N e u p e r , K .-R.
Müller , and D. Mattia . 2010 Combining brain-computer interfaces and
assistive technologies: state-of-the-art and challenges . Front Neurosci 4 : 161 .
Miner , L. A. , D. J. McFarland , and J. R. Wolpaw . 1998 . Answering questions
with an electroencephalogram-based brain-computer interface . Arch Phys
Med Rehabil 79 ( 9 ): 1029 – 1033 .
M i z u t a n i , T . , M . A k i , R . S h i o z a w a , M . U n a k a m i , T . N o z a w a , K . Ya j i m a , H.
Tanabe , and M. Hara . 1990 . Development of ophthalmoplegia in amyo-
trophic lateral sclerosis during long-term use of respirators . J Neurol Sci
99 ( 2–3 ): 311 – 319 .
Mugler , E.M. , C. A. Ruf , S. Halder , M. Bensch , and A. Kubler . 2010 . Design and
implementation of a P300-based brain-computer interface for controlling
an internet browser . IEEE Trans Neural Sys & Rehabil Eng 18 ( 6 ): 599 – 609 .
Muller-Putz , G. R. , R. Scherer , G. Pfurtscheller , and R. Rupp . 2005 . EEG-based
neuroprosthesis control: a step towards clinical practice . Neurosci Lett
382 ( 1–2 ): 169 – 174 .
Münßinger , J. I. , S. Halder , S. C. Kleih , A. Furdea , V. Raco , A. Hösle , and A.
Kübler . 2010 . Brain painting:  rst evaluation of a new brain–computer
interface application with ALS-patients and healthy volunteers . Front
Neurosci 4 : 182 , ( http://www.frontiersin.org/neuroprosthetics/10.3389/
fnins.2010.00182/full )
Murray B ., Natural history and prognosis in amyotrophic lateral sclerosis. In
H. Mitsomoto , S. Przedborski , and P. H. Gordon (Eds.), Amyotrophic Lat-
eral Sclerosis . New York : Taylor & Francis , pp. 227 – 255 .
N a i t o , M . , Y. M i c h i o k a , K . O z a w a , Y . I t o , M . K i g u c h i , a n d T . K a n a z a w a . 2007 .
A communication means for totally locked-in ALS patients based on
changes in cerebral blood volume measured with near-infrared light .
IEICE Trans Inf Syst E90-D ( 7 ): 1028 – 1036 .
20-Wolpaw-20.indd 335
20-Wolpaw-20.indd 335
8/10/2011 5:14:42 PM
8/10/2011 5:14:42 PM
OUP UNCORRECTED PROOF – FIRST-PROOF, 11/08/2011, GLYPH

336 | BRAIN–COMPUTER INTERFACES
122
121
120
119
118
117
116
115
114
113
112
111
110
109
108
107
106
105
104
103
102
101
100
99
98
97
96
95
94
93
92
91
90
89
88
87
86
85
84
83
82
81
80
79
78
77
76
75
74
73
72
71
70
69
68
67
66
65
64
63
62
61
60
59
58
57
56
55
54
53
52
51
50
49
48
47
46
45
44
43
42
41
40
39
38
37
36
35
34
33
32
31
30
29
28
27
26
25
24
23
22
21
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
National ALS Registry Home Page . https://wwwn.cdc.gov/ALS/ALSResourc-
es.aspx .
Neumann , N. and A. Kubler . 2003 . Training locked-in patients: a challenge for
the use of brain-computer interfaces . IEEE Trans Neural Syst Rehabil Eng
11 ( 2 ): 169 – 172 .
N i j b o e r , F . , A . F u r d e a , I . G u n s t , J . M e l l i n g e r , D . J . M c F a r l a n d , N . B irbaumer ,
and A. Kubler . 2008 . An auditory brain-computer interface (BCI) . J
Neurosci Methods 167 ( 1 ): 43 – 50 .
N i j b o e r , F. , E . W. S e l l e r s , J . M e l l i n g e r , M . A . J o r d a n , T . M a t u z , A . Furdea , S. Halder ,
U . M o c h t y , D . J . K r u s i e n s k i , T . M . V a u g h a n , J . R . W o l p a w , N . B i r b a u m e r ,
and A. Kubler . 2008 . A P300-based brain-computer interface for people with
amyotrophic lateral sclerosis . Clin Neurophysiol 119 ( 8 ): 1909 – 1916 .
Nygren , I. , and H. Askmark . 2006 . Self-reported quality of life in amyotrophic
lateral sclerosis . J.Palliative Med 9 : 304 – 308 .
Pedersen , P.M. , H.S. J ø rgensen , H. Nakayama , H. O. Raaschou , and T. S. Olsen .
1995 Aphasia in acute stroke: incidence, determinants, and recovery . Ann
Neurol 38 : 659 – 666 .
Pfurtscheller , G. , C. Guger , G. Muller , G. Krausz , and C. Neuper . 2000 . Brain oscil-
lations control hand orthosis in a tetraplegic . Neurosci Lett 292 ( 3 ): 211 – 214 .
Phillips , W. , A. Anderson , M. Rosengreen , J. Johnson , and J. Halpin . 2010 .
Monitoring sedation status over time in ICU patients: reliability and valid-
ity of the Richmond Agitation-Sedation Scale (RASS) . J Pain Palliat Care
Pharmacother 24 ( 4 ): 349 – 355 .
P i c c i o n e , F. , F . G i o r g i , P. T o n i n , K . P r i  i s , S . G i o v e , S . S i l v o n i , G . P a l m a s , a n d F .
Beverina . 2006 . P300-based brain computer interface: reliability and
performance in healthy and paralysed participants . Clin Neurophysiol
117 ( 3 ): 531 – 537 .
Pinto , S. , and M. de Carvalho . 2008 . Amyotrophic lateral sclerosis patients and
o c u l a r p t o s i s . Clin Neurol Neurosurg 110 ( 2 ): 168 – 170 .
Pires , G. , U. Nunes , and M. Castelo-Branco . 2011 . Statistical spatial  ltering for
a P300-based BCI: tests in able-bodied, and patients with cerebral palsy
and amyotrophic lateral sclerosis . J Neurosci Methods 195 ( 2 ): 270 – 281 .
Popescu , F. , S. Fazli , Y. Badower , B. Blankertz and K.-R. Müller . 2007 . S i n g l e
trial classi cation of motor imagination using 6 dry EEG electrodes . PLoS
One 2 : e637 .
R o b b i n s , R . A . , Z . S i m m o n s , B . A . B r e m e r , S . M . W a l s h , a n d S . F i s c h e r . 2 0 01 .
Quality of life in ALS is maintained as physical function declines . Neurology
56 ( 4 ): 442 – 444 .
Ryan , D.B. , G.E. Frye , G. Townsend , D.R. Berry , S. Mesa-G. , N.A. Gates , and
E.W . Sellers . 2011 . Predictive spelling with a P300-based brain-computer
interface: Increasing the rate of communication . Int J Hum Comput Interact
27(1):69–84.
S c h a l k , G . , D . J . M c F a r l a n d , T . H i n t e r b e r g e r , N . B i r b a u m e r a n d J . R . W o l p a w .
2004 . BCI2000: a general-purpose brain-computer interface (BCI) system .
IEEE Trans Biomed Eng 51 : 1034 – 1043 .
Scherer , M. J. , G. Craddock , and T. Mackeogh . 2011 .  e relationship of per-
sonal factors and subjective well-being to the use of assistive technology
devices . Disabil Rehabil 33 ( 10 ): 811 – 817 . Epub 2010 Aug 24 .
Schreuder , M. , B. Blankertz , and M. Tangermann . 2010 . A new auditory multi-
class brain-computer interface paradigm: spatial hearing as an informative
c u e . PLoS One 5 ( 4 ): e9813 .
Sellers , E. W. , and E. Donchin . 2006 . A P300-based brain-computer interface:
initial tests by ALS patients .
Clin Neurophysiol 117 ( 3 ): 538 – 548 .
S e l l e r s , E . W. , D . J . K r u s i e n s k i , D . J . M c F a r l a n d , T . M . V a u g h a n , a n d J . R .
Wolpaw . 2006 . A P300 event-related potential brain-computer interface
(BCI): the e ects of matrix size and inter stimulus interval on performance .
Biol Psychol 73 ( 3 ): 242 – 252 .
Sellers , E. W. , G. Schalk , and M. Donchin . 2003 .  e P300 as a typing tool: tests of
brain computer interface with an ALS patient . Psychophysiology 42 ( S1 ): S29 .
S e l l e r s , E . W. , P . T u r n e r , W. A . S a r n a c k i , T . M c m a n u s , T. M . V a u g h a n , a n d R .
Matthews , R . 2009 . A novel dry electrode for brain–computer interface .
In Proceedings of the 13th International Conference on Human-Computer
Interaction. Part II . Berlin : Springer-Verlag , pp. 623 – 631 .
Sellers , E. W. , T. M. Vaughan and J. R. Wolpaw . 2010 . A brain-computer
interface for long-term independent home use . Amyotroph Lat Scler 11 ( 5 ):
449 – 455 .
Shields , A. M. , M. Park , S. E. Ward , and M. K. Song . 2010 . Subject recruitment
and retention against quadruple challenges in an intervention trial of end-
of-life communication . J Hosp Palliat Nurs 12 ( 5 ): 312 – 318 .
S i l v o n i , S . , C . V o l p a t o , M . C a v i n a t o , M . M a r c h e t t i , K . P r i  i s , A . M e r i c o , P.
Tonin , K. Koutsikos , F. Beverina , and F. Piccione . 2009 . P300-based brain-
computer interface communication: evaluation and follow-up in amyo-
trophic lateral sclerosis . Front Neurosci 3 : 60 . ( http://www.frontiersin.org/
neuroprosthetics/10.3389/neuro.20.001.2009/full )
S i m m o n s Z . , S . H . F e l g o i s e , B . A . B r e m e r , S . M . Wa l s h , D . J . H u  o r d , M . B .
B r o m b e r g , W. D a v i d , D . A . F o r s h e w , T . D . H e i m a n - P a t t e r s o n , E . C . L a i , a n d
L. McCluskey . 2006 .  e ALSSQOL: balancing physical and nonphysical
factors in assessing quality of life in ALS . Neurology 67 ( 9 ): 1659 – 1664 .
S t r e i  , E. B . 1967 . Gerontology and geriatrics of the eye . Surv Ophthalmol
12 ( 4 ): 311 – 323 .
Tatum , W. O . 2001 . Long-term EEG monitoring: a clinical approach to electro-
physiology . J Clin Neurophysiol 18(5) : 442 – 455 .
Tecchio , F. , C. Porcaro , G. Barbati , and F. Zappasodi . 2007 . Functional source
separation and hand cortical representation for a brain-computer interface
feature extraction . J Physiol 580 ( Pt. 3 ): 703 – 721 .
Towler , M. L. , B. D. Beall , and J. B. King . 1962 . Drug e ects on the electroen-
cephalographic pattern, with speci c consideration of diazepam . South
Med J 55 : 832 – 838 .
T o w n s e n d , G . , B . K . L a P a l l o , C . B . B o u l a y , D . J . K r u s i e n s k i , G . E . F r y e , C. K.
H a u s e r , N . E . S c h w a r t z , T. M . V a u g h a n , J . R . W o l p a w , a n d E . W. S e l l e r s .
2010 . A novel P300-based brain-computer interface stimulus presentation
paradigm: moving beyond rows and columns . Clin Neurophysiol
121 ( 7 ): 1109 – 1120 .
Tucker , J. , and W. N. Charman . 1975 .  e depth-of-focus of the human eye for
Snellen letters . Am J Optom Physiol Opt 52 ( 1 ): 3 – 21 .
van Gerven , M. , and O. Jensen . 2009 . Attention modulations of posterior alpha
as a control signal for two-dimensional brain-computer interfaces . J
Neurosci Methods 179 ( 1 ): 78 – 84 .
V a u g h a n , T . M . , D . J . M c F a r l a n d , G . S c h a l k , W. A . S a r n a c k i , D . J . K r u s i e n s k i , E .
W. Sellers , and J. R. Wolpaw . 2006 .  e Wadsworth BCI Research and
Development Program: at home with BCI . IEEE Trans Neural Syst Rehabil
Eng 14 ( 2 ): 229 – 233 .
V o l p a t o , C . , F . P i c c i o n e , S . S i l v o n i , M . C a v i n a t o , A . P a l m i e r i , F . M e neghello ,
and N. Birbaumer . 2010 . Working memory in amyotrophic lateral sclero-
sis: auditory event-related potentials and neuropsychological evidence . J
Clin Neurophysiol 27 ( 3 ): 198 – 206 .
Wade , D. T. , R. L. Hewer , R. M. David RM and P. M. Menderby . 1986 Aphasia
a er stroke: natural history and associated de cits . J Neurol Neurosurg
Psychiartry 49 : 11 – 16 .
Wilkins , V. M. , M. L. Bruce , and J. A. Sirey . 2009 . Caregiving tasks and training
interest of family caregivers of medically ill homebound older adults . J
Aging Health 21 ( 3 ): 528 – 542 .
Williamson , J. , R. Murray-Smith , B. Blankertz , M. Krauledat , and K.-R.
Mueller . 2009 . Designing for uncertain, asymmetric control: Interactions
design for brain-computer interface . Int J Hum Comput Stud 67 ( 10 ): 827 –
841 .
Wolf , C . 2011 Personal communication . Woolley , S. C. , M. K. York , D. H.
Moore , A. M. Strutt , J. Murphy , P. E. Schulz , and J. S. Katz . 2010 . Detecting
frontotemporal dysfunction in ALS: utility of the ALS Cognitive Behavioral
Screen (ALS-CBS) . Amyotroph Lat Scler 11 ( 3 ): 303 – 311 .
Young , G. B. , and V. C. Campbell . 1999 . EEG monitoring in the intensive care
unit: pitfalls and caveats . J Clin Neurophysiol 16 ( 1 ): 40 – 45 .
20-Wolpaw-20.indd 336
20-Wolpaw-20.indd 336
8/10/2011 5:14:43 PM
8/10/2011 5:14:43 PM
OUP UNCORRECTED PROOF – FIRST-PROOF, 11/08/2011, GLYPH