Academic Edi o : Jona han Fos e
Recei ed: 8 No embe 2024
Re ised: 9 Decembe 2024
Accep ed: 1 Janua y 2025
Published: 2 Janua y 2025
Ci a ion: Sen a, C.C.; Sampaio,
A.C.S.; Lapen a, O.M. Visual Noise
Mask o Human Poin -Ligh Displays:
A Coding-F ee App oach. Neu oSci
2025,6, 2. h ps://doi.o g/10.3390/
neu osci6010002
Copy igh : © 2025 by he au ho s.
Licensee MDPI, Basel, Swi ze land.
This a icle is an open access a icle
dis ibu ed unde he e ms and
condi ions o he C ea i e Commons
A ibu ion (CC BY) license
(h ps://c ea i ecommons.o g/
licenses/by/4.0/).
A icle
Visual Noise Mask o Human Poin -Ligh Displays:
A Coding-F ee App oach
Ca a ina Ca alho Sen a , Ad iana Conceição Soa es Sampaio and Oli ia Mo gan Lapen a *
Psychological Neu oscience Labo a o y, Psychology Resea ch Cen e , School o Psychology, Uni e si y o Minho,
Rua da Uni e sidade, 4710-057 B aga, Po ugal; [email p o ec ed] (C.C.S.);
[email p o ec ed] (A.C.S.S.)
*Co espondence: [email p o ec ed]
Abs ac : Human poin -ligh displays consis o luminous do s ep esen ing human a ic-
ula ions, hus depic ing ac ions wi hou pic o ial in o ma ion. These s imuli a e widely
used in ac ion ecogni ion expe imen s. Because humans excel in decoding human mo ion,
poin -ligh displays (PLDs) a e o en masked wi h addi ional mo ing do s (noise masks),
he eby challenging s imulus ecogni ion. These noise masks a e ypically ound wi hin
p op ie a y p og amming so wa e, en ail ile o ma es ic ions, and demand ex ensi e
p og amming skills. To add ess hese limi a ions, we p esen he i s use - iendly s ep-
by-s ep guide o de elop isual noise o mask PLDs using ee, open-sou ce so wa e ha
o e s compa ibili y wi h a ious ile o ma s, ea u es a g aphical in e ace, and acili a es
he manipula ion o bo h 2D and 3D ideos. Fu he , o alida e ou app oach, we es ed
wo gene a ed masks in a pilo expe imen wi h 12 subjec s and demons a ed ha hey
e ec i ely jeopa dised human agen ecogni ion and, he e o e, ac ion isibili y. In sum, he
main ad an ages o he p esen ed me hodology a e i s cos -e ec i eness and ease o use,
making i appealing o no ices in p og amming. This ad ancemen holds he po en ial o
s imula e young esea che s’ use o PLDs, os e ing u he explo a ion and unde s anding
o human mo ion pe cep ion.
Keywo ds: poin -ligh walke s; isual noise; ac ion pe cep ion; Blende so wa e
1. In oduc ion
Poin -ligh displays (PLDs) a e widely used s imuli o in es iga e human mo emen
ecogni ion and p ocessing. Since he pionee ing wo k o Johansson in 1973 [
1
], his
echnique has been highly applied in se e al domains, including biomechanics, mo o , and
cogni i e neu oscience [
2
]. These s imuli a e gene a ed by ilming human ac o s wea ing
da k sui s wi h e lec i e pa ches a hei majo join s [
1
,
3
,
4
]. Ac ion unde s anding is
c ucial o human in e ac ions [5].
The neu al p ocessing pa hway o poin -ligh ac ions in ol es mul iple s ages, om
isual inpu o mo ion con ol. Ini ially, he eyes p ocess and ansmi he isual in o ma ion
o he p ima y isual co ex (V1) [
6
]. F om he e, he in o ma ion lows h ough he do sal
and en al isual s eams, whe e he i s iden i ica ion o space posi ion and he isual
elemen s’ ca ego isa ion occu [
7
]. Subsequen ly, he supe io empo al gy us and in e io
pa ie al lobe a e ac i a ed [
8
], as a e he en al po ion o he p emo o co ex and in e io
on al gy us [
6
]. These a eas a e pa o he mi o -neu on sys em and a e c ucial o ac ion
unde s anding [
9
]. This ac i a ion indica es ha e en simpli ied poin -ligh anima ions
can ec ui ac ion obse a ion ne wo ks. In ac , human PLDs ha e been shown o e oke
Neu oSci 2025,6, 2 h ps://doi.o g/10.3390/neu osci6010002
Neu oSci 2025,6, 2 2 o 11
simila co icospinal exci abili y o eal human mo emen , sugges ing hei e ec i eness in
mo o esonance and ac ion obse a ion asks [10].
Impo an ly, his p ocess is no s ic ly linea . The e a e in e ac ions be ween he do sal
and en al s eams [
11
], and in o ma ion p ocessing occu s in pa allel ac oss mul iple
b ain egions, especially in ela ion o highe -le el p ocessing du ing ac ion iden i ica ion.
In sum, he neu al p ocessing o poin -ligh ac ions in ol es a dis ibu ed ne wo k
o b ain egions, om ea ly isual a eas o highe -le el egions esponsible o ac ion
unde s anding, ep esen a ion, and an icipa ion. This pa hway demons a es he b ain’s
abili y o ex ac meaning ul in o ma ion om spa se isual inpu s and link i o mo o
ep esen a ions and ac ion unde s anding.
Conside ing ha humans a e expe s in ecognising and p edic ing human mo ion,
Cu ing and colleagues [
12
] sugges ed adding isual noise o mask PLD s imuli, hus
inc easing ask complexi y. They es ed s a ic, dynamic, and linea masks. The s a ic mask
consis ed o addi ional s a iona y do s; o he dynamic mask, he noise do s andomly
changed on each ame; and he linea mask consis ed o he do elemen s mo ing in a
speci ic o ien a ion [
12
]. The s a ic and linea masks we e ine ec i e in dis u bing he iden-
i ica ion o he human agen and i s ac ion compa ed o he no-mask condi ion. In con as ,
he dynamic mask dec eased pe o mance compa ed o all o he condi ions, making i
ha de o pa icipan s o iden i y he human agen and i s mo emen . Un o una ely, his
epo p o ides li le in o ma ion on how o cons uc masks and only p esen s gene al
concep s o mo emen ype and applied eloci y [12].
Subsequen s udies ha e adop ed masking s a egies using addi ional mo ing do s
(e.g., [
13
,
14
]. Deca oi e and colleagues [
9
] de eloped PLAViMoP (Poin -Ligh Display
Visualisa ion and Modi ica ion Pla o m), a so wa e con aining wo e sions (1 and 2) o
isualising and ans o ming 3D poin -ligh sequences. The pla o m o e s an in e es ing
app oach ha allows use s o add ce ain ypes o masks o hei PLD sequences; howe e ,
i has some cons ain s, such as equi ing speci ic so wa e and ile o ma s [15].
Resea ch cons an ly needs di e se s imuli ma e ial [
16
], and hus, se e al echniques
ha e been p oposed o c ea e PLD s imuli and masks [
3
,
16
–
18
]. S ill, he e is a lack o
guidelines, as mos a ailable ma e ials do no p o ide s ep-by-s ep explana ions and
some imes do no speci y he so wa e used [
12
,
19
], which hampe s he consis ency and
ep oducibili y o noise mask de elopmen . Fu he mo e, mos p ocedu es demand speci ic
and paid so wa e ha usually suppo s speci ic ile o ma s used in biomechanics [
15
].
Addi ionally, hey equi e use s o ha e p og amming skills i he e is a need o cus omise
he mask o he desi ed ou come [15].
To add ess hese issues, we desc ibe a s ep-by-s ep p ocess o c ea ing and adap ing
isual noise masks o PLDs using ee, use - iendly so wa e ha equi es no p og am-
ming expe ience, as i can be ope a ed solely ia a g aphical in e ace. This me hod aims o
acili a e he c ea ion o PLD s imuli and masks o esea che s and s uden s a any ca ee
s age, making he p ocess much less ime-consuming and o e ing a as e lea ning p ocess
han p og amming. E en in in o ma ics echnology, no ice p og amme s ace a long- e m
lea ning cu e be o e ha ing a solid g asp on his subjec [20].
To achie e ou goals, we selec ed a ee so wa e commonly used o game de el-
opmen [
21
] ha allows playing wi h many p ope ies ha o he s do no . Speci ically,
i suppo s a ious ile o ma s and allows cus omisa ion o ame a e, ex u e, colou ,
and opaci y. I o e s a g aphical in e ace, making i accessible wi hou p e equisi es. In
sum, he majo ad an ages iden i ied we e (a) he simplici y and accessibili y o he so -
wa e i sel , (b) he ease o modi ying and manipula ing he pa ame e s o he masks, and
(c) he possibili y o c ea e he s imuli om sc a ch. The e o e, we de ail he cons uc ion
o wo PLD noise masks wi h 22 and 33 do s as well as how o me ge hem wi h PLDs
Neu oSci 2025,6, 2 3 o 11
and, inally, p esen he esul s om a pilo expe imen es ing i s e icacy on challenging
s imuli ecogni ion.
2. Ma e ials and Me hods
2.1. S ep-by-S ep P ocedu e o Building he Noise Mask
Fo c ea ing ou s imuli, we used Blende so wa e 2.91.2 (Blende Founda ion and
Ins i u e, Ams e dam, The Ne he lands [
21
]), which suppo s a wide ange o ile o ma s,
such as OBJ, FBX, 3DS, PLY, and STL, o be impo ed and expo ed.
Once in he so wa e en i onmen (a ailable o download a h ps://www.blende .
o g, accessed on 31 Decembe 2024), he i s s ep is o choose ei he he 2D o 3D op ion
acco ding o speci ic needs and objec i es. No e ha he so wa e suppo s he en i e
3D pipeline, including modelling, igging, simula ion, ende ing, composi ing, mo ion
acking, and ideo edi ing [
21
]. The 2D anima ion pipeline also includes ea u es such as
colou ing, anima ing, applying modi ie s, cu -ou anima ion, mo ion g aphics, and g ease
pencil ools [21].
Ou p ojec used Blende ’s 2D ea u e o align wi h he biological PLD da a om
Lapen a and colleagues’ [
4
] da abase. Please no e ha mos 2D ins uc ions can also be
applied when using 3D op ions. Finally, we used he Blende edi ing heade menu o
o e lay he human PLD s imuli wi h he newly c ea ed mask.
Nex , we selec ed he D aw Mode o es ablish he desi ed pa ame e s o he masking
do s. We hen chose he D awing a Ci cle op ion, d ew wi h he mouse, and p essed he
‘En e ’ key once inished. Then, we swi ched o Objec Mode o access a oolba con aining
ools o check he d awn objec ’s measu emen s and se i s displacemen along he ‘x’ and
‘y’ axes.
In he igh panel ( he P ope ies Edi o ), we se he backg ound colou o black by
clicking on Wo ld P ope ies, which enables backg ound se ings. We also selec ed he s oke
colou as whi e unde Ma e ial P ope ies, whe e he base colou , shape, o mode o any
d awing gene a ed o e he backg ound can be de ined. Addi ionally, we speci ied he
s oke ex u e as solid, wi h opaci y se o 1.
Finally, we assigned hal o he do s a size o 0.100965 m and he o he hal a size
o 0.060999 m. These sizes we e de e mined h ough ial and e o . Ini ially, he do s in
he o iginal s imuli [
4
] we e measu ed o de e mine he a e age size o he isible poin s.
Based on hese measu emen s, he sizes o he mask do s we e adjus ed by es ing sligh ly
smalle and la ge alues un il he desi ed isual con igu a ion was achie ed. We chose
hese wo sizes o e lec sligh a ia ions in do size wi hin and be ween he do s o ou
PLD s imuli. These a ia ions occu due o he agen ’s mo emen in space, as he PLD was
c ea ed using human pa icipan s wea ing e lec i e pa ches [
4
]. Mino changes in do size
a ise based on he dis ance om he came a: o ins ance, in a walking mo emen , do s
g adually inc ease in size as he agen mo es close o he came a. Simila size changes
occu in mo emen s in ol ing he legs o a ms, such as a low kick o o ea m mo ion. Once
he colou , opaci y, size, and ex u e we e se , he do s we e copied in he Scene Anima ion
panel o o ganise he quan i y o do s and hei ini ial posi ions on he sc een, speci ically
he displacemen o 22 o 33 do s.
Once he do s we e eady, we c ea ed hei dynamic mo emen , a key p ope y ha ,
in a 2D en i onmen , elies on adjus ing an objec ’s posi ion along he ‘x’ and ‘y’ axes. To
change a do ’s posi ion, we accessed he oolba , selec ed he Mo e op ion, and adjus ed he
do along he ‘x’ and ‘y’ axes. Fo smoo h mo emen , i is essen ial o change he posi ion
g adually, ei he on e e y ame o ac oss a sequence o ames, depending on he o iginal
PLD ideos and he desi ed speed o he mask. Fo ou masks, he ideal speed o ma ch
ou s imuli was de ined as a displacemen e e y 7 ames.
Neu oSci 2025,6, 2 4 o 11
To se hese pa ame e s, we swi ched be ween he Objec and Edi modes o clicked he
Edi bu on a ailable o each objec on he iew laye (g een bu on). A help ul p ac ice o
p e en do s om swapping posi ions ins ead o mo ing smoo hly is o use he In e pola e
Sequence command a he op o he wo kspace. This command should be applied a e e y
ame in e al and o each do .
Simila o he size o he do s, he mo emen in e als we e es ablished by obse ing
he mo ion dynamics in he o iginal ideos, ensu ing ha he mask do s blended na u ally
wi h he ideo eloci y while a oiding o e lap o eplica ion o he o iginal mo emen . The
mask aimed o c ea e enough dis up ion o make ecognising he mo ion mo e challenging.
This app oach was p ima ily guided by quali a i e analysis, wi h he goal o balancing
in e e ence in he s imulus.
I is ad isable o sa e p og ess a each s ep using he Sa e As op ion, which ou pu s a
Blende p ojec ile necessa y o ende ing he inal ideo. To ende he ideo, we accessed
he Rende ing menu, selec ed he Rende Anima ion op ion, and con igu ed se ings such as
he desi ed ames pe second (see Figu e 1 o a diag am o hese s eps).
Neu oSci 2025, 6, x FOR PEER REVIEW 4 o 11
To se hese pa ame e s, we swi ched be ween he Objec and Edi modes o clicked
he Edi bu on a ailable o each objec on he iew laye (g een bu on). A help ul p ac-
ice o p e en do s om swapping posi ions ins ead o mo ing smoo hly is o use he
In e pola e Sequence command a he op o he wo kspace. This command should be ap-
plied a e e y ame in e al and o each do .
Simila o he size o he do s, he mo emen in e als we e es ablished by obse ing
he mo ion dynamics in he o iginal ideos, ensu ing ha he mask do s blended na u ally
wi h he ideo eloci y while a oiding o e lap o eplica ion o he o iginal mo emen .
The mask aimed o c ea e enough dis up ion o make ecognising he mo ion mo e chal-
lenging. This app oach was p ima ily guided by quali a i e analysis, wi h he goal o bal-
ancing in e e ence in he s imulus.
I is ad isable o sa e p og ess a each s ep using he Sa e As op ion, which ou pu s
a Blende p ojec ile necessa y o ende ing he inal ideo. To ende he ideo, we ac-
cessed he Rende ing menu, selec ed he Rende Anima ion op ion, and con igu ed se ings
such as he desi ed ames pe second (see Figu e 1 o a diag am o hese s eps).
The iles o he masks desc ibed he e and he Blende p ojec a e a ailable o down-
load as .zip a chi es in an online eposi o y (h ps://os .io/q2u95/, accessed on 31 Decem-
be 2024).
Figu e 1. Diag am o he mask-building p ocess. The diag am p o ides an o e iew o he s ep-by-
s ep p ocedu e, highligh ing key s ages om opening he so wa e o ende ing he ideo.
Figu e 1. Diag am o he mask-building p ocess. The diag am p o ides an o e iew o he s ep-by-
s ep p ocedu e, highligh ing key s ages om opening he so wa e o ende ing he ideo.
The iles o he masks desc ibed he e and he Blende p ojec a e a ailable o down-
load as .zip a chi es in an online eposi o y (h ps://os .io/q2u95/, accessed on 31 Decem-
be 2024).
2.2. Combining Videos wi h he Mask
Once he noise masks we e p epa ed, we combined each o hem wi h ou PLD s imuli
using he same so wa e. The e o e, we used he Video Edi ing ool; by using his selec ion,
he desi ed ideo and he mask only ha e o be d agged o he edi ion panel. Each ideo
Neu oSci 2025,6, 2 5 o 11
was selec ed indi idually, and we hen added an E ec S ip wi h he T ans o m op ion (one
o se e al a ailable op ions). This allowed each mask and s imuli ideo o ha e a ans o m
s ip a ached o i , enabling he in eg a ion o he mask ideo in o he s imuli ideo.
Impo an ly, he o iginal PLD did no unde go any enhancemen o modi ica ion o
i s o iginal p ope ies. The PLD s imuli e ained hei inhe en cha ac e is ics, such as do
sizes and b igh ness, wi hou any al e a ions beyond he addi ion o he noise mask.
Nex , we selec ed he Composi ing op ion (unde he Ac i e Tool op ion), which con ains
he Blende and Opaci y p ope ies necessa y o adjus he pa ame e s o combining he
ideos and c ea ing he o e lapping e ec . Fo he mask ideo, we selec ed he Alpha O e
blending op ion and se he opaci y o 1. Fo he o e lapping, we selec ed he blende
op ion o Alpha O e o he mask ideos and se he opaci y o 1. Fo he s imuli ideos,
we selec ed he blende op ion o Add, and he opaci y was also se o 1.
These op ions we e chosen o he mask and s imuli because Alpha O e de ines how
o laye he mask on op o he ideo, while he Add op ion speci ies which ideo will
ecei e he mask. The opaci y se ing con ols he anspa ency o he ideos; ou selec ion
ensu ed ha bo h he mask and s imuli ideos we e ully opaque wi h no anspa ency.
Once his p ocess was comple e, he ideo was ende ed as desc ibed p e iously (see
Figu e 2 o an illus a ion o he inal ideos).
Neu oSci 2025, 6, x FOR PEER REVIEW 5 o 11
2.2. Combining Videos wi h he Mask
Once he noise masks we e p epa ed, we combined each o hem wi h ou PLD s im-
uli using he same so wa e. The e o e, we used he Video Edi ing ool; by using his selec-
ion, he desi ed ideo and he mask only ha e o be d agged o he edi ion panel. Each
ideo was selec ed indi idually, and we hen added an Effec S ip wi h he T ans o m op-
ion (one o se e al a ailable op ions). This allowed each mask and s imuli ideo o ha e
a ans o m s ip a ached o i , enabling he in eg a ion o he mask ideo in o he s imuli
ideo.
Impo an ly, he o iginal PLD did no unde go any enhancemen o modi ica ion o
i s o iginal p ope ies. The PLD s imuli e ained hei inhe en cha ac e is ics, such as do
sizes and b igh ness, wi hou any al e a ions beyond he addi ion o he noise mask.
Nex , we selec ed he Composi ing op ion (unde he Ac i e Tool op ion), which con-
ains he Blende and Opaci y p ope ies necessa y o adjus he pa ame e s o combining
he ideos and c ea ing he o e lapping effec . Fo he mask ideo, we selec ed he Alpha
O e blending op ion and se he opaci y o 1. Fo he o e lapping, we selec ed he blende
op ion o Alpha O e o he mask ideos and se he opaci y o 1. Fo he s imuli ideos,
we selec ed he blende op ion o Add, and he opaci y was also se o 1.
These op ions we e chosen o he mask and s imuli because Alpha O e de ines how
o laye he mask on op o he ideo, while he Add op ion speci ies which ideo will
ecei e he mask. The opaci y se ing con ols he anspa ency o he ideos; ou selec ion
ensu ed ha bo h he mask and s imuli ideos we e ully opaque wi h no anspa ency.
Once his p ocess was comple e, he ideo was ende ed as desc ibed p e iously (see
Figu e 2 o an illus a ion o he inal ideos).
Figu e 2. Examples o biological PLDs, speci ically ame 15 o he jumping jack ac ion, in he h ee
masking condi ions: NoMask, Mask22, and Mask33. The op ow illus a ions show he ideo as
p esen ed in he ask. In con as , he bo om ow depic s he same images bu wi h he do s con-
nec ed o allow isualisa ion o he human agen .
2.3. Tes ing he Masks
Once he ideos we e eady, hey we e es ed o e alua e whe he , as expec ed, add-
ing dynamic masking wi h 22 and 33 do s o he PLD would inc ease he difficul y o
s imuli de ec ion. The e o e, we conduc ed a pilo expe imen wi h 12 pa icipan s (6
male) aged 22–39 yea s (M = 33.42; SD = 5.18). Simila sample sizes ha e been adop ed in
Figu e 2. Examples o biological PLDs, speci ically ame 15 o he jumping jack ac ion, in he h ee
masking condi ions: NoMask, Mask22, and Mask33. The op ow illus a ions show he ideo as
p esen ed in he ask. In con as , he bo om ow depic s he same images bu wi h he do s connec ed
o allow isualisa ion o he human agen .
2.3. Tes ing he Masks
Once he ideos we e eady, hey we e es ed o e alua e whe he , as expec ed, adding
dynamic masking wi h 22 and 33 do s o he PLD would inc ease he di icul y o s imuli
de ec ion. The e o e, we conduc ed a pilo expe imen wi h 12 pa icipan s (6 male) aged
22–39 yea s (M = 33.42; SD = 5.18). Simila sample sizes ha e been adop ed in expe imen s
using his ype o s imuli in ac ion pe cep ion s udies (e.g., [
6
,
22
,
23
]). The expe imen
ollowed he Decla a ion o Helsinki and was app o ed by he local E hics Commi ee o
In es iga ion on Social and Human Science o he Uni e si y o Minho, B aga, Po ugal
(CEICSH 069/2020). Pa icipan s we e ec ui ed using a con enience sampling me hod
h ough a ne wo k-based app oach. The sample consis ed o indi iduals om a ious
Neu oSci 2025,6, 2 6 o 11
p o essional backg ounds, including a pilo , a e e ina ian, a ashion designe , a che , wo
psychologis s, wo economis s, a biologis , a ou ism p o essional, a philosophe , and a
hospi ali y p o essional. This di e si y ensu ed ha he pa icipan s we e no solely om
a psychology backg ound, he eby minimising he isk o bias ela ed o amilia i y wi h
PLD s imuli. All pa icipan s ga e hei w i en in o med consen and complied wi h he
ollowing c i e ia: age be ween 18 and 45 yea s, no mal o co ec ed- o-no mal isual
acui y, no known pas o cu en psychological o psychia ic diso de s, and no his o y o
cen al ne ous sys em inju y.
The selec ed subse o 7 biological (Bio) and 7 sc ambled (Sc ) mo emen s om
Lapen a and colleagues [
4
], combined wi h he c ea ed masks ( o de ails on he s imuli
and he me hods used o hei c ea ion and alida ion, please e e o [
4
]), we e used in he
expe imen . Th ee condi ions we e gene a ed o each s imuli ype (Bio s. Sc ), speci ically
NoMask (o iginal s imuli wi hou isual noise), Mask22 (o iginal s imuli + 22 do s mask),
and Mask33 (o iginal s imuli + 33 do s mask).
The ask was buil and p esen ed using PsychoPy so wa e ( 2022.2.4) [
24
] and com-
p ised a aining phase wi h 6 ials (one o each s imuli condi ion combina ion) o amil-
ia ise pa icipan s wi h he ask and cla i y any doub s. Then, h ee expe imen al blocks
we e p esen ed.
The i s block (B1) con ained 28 s imuli (7 Bio and 7 Sc om he Mask22 and Mask33
condi ions), he second block (B2) con ained he 14 NoMask s imuli, and he hi d block
(B3) included all 42 ideos. We chose his app oach because mos s udies p esen only
masked s imuli, and we wan ed o in es iga e whe he knowing he unmasked s imuli
(i.e., a e he NoMask condi ion) would a ec de ec ion. S imuli we e p esen ed andomly
wi hin each block. Each ial began wi h a g ey sc een displaying a black ixa ion c oss
(400–900 ms), ollowed by he ideo p esen a ion (~1000 ms), and so o h. Pa icipan s
we e ins uc ed o espond as as and accu a ely as possible du ing he ideo p esen a ion,
p essing one o wo keys on he keyboa d o indica e whe he o no hey de ec ed a human
ac ion in he s imulus. The esponse was de ined as accu a e i pa icipan s de ined a Bio
ideo as con aining human ac ion and a Sc ideo as no con aining human ac ion.
3. Resul s
Ou ini ial esul was he gene a ion o wo ypes o dynamic masks, one con ain-
ing 22 do s and ano he wi h 33 do s, whe eas he second ou come consis ed o es ing
hem. These masks we e subsequen ly applied o selec ed poin -ligh s imuli, c ea ing
six expe imen al s imulus ca ego ies—BioNoMask, BioMask22, BioMask33, Sc NoMask,
Sc Mask22, and Sc Mask33— o es hei e icacy. The esul s om hese es s cons i u ed
he second ou come.
To es he e icacy o hese masks in obs uc ing s imulus iden i ica ion, we i s
examined whe he he p esen a ion o unmasked s imuli a ec ed he iden i ica ion o
masked s imuli. To his pu pose, we conduc ed epea ed measu es analyses o a iance
( mANOVA) o accu acy and eac ion ime o (i) unmasked s imuli (NoMask) iden i ica-
ion, wi h block (B2 s. B3) and s imuli ype (Bio s. Sc ) as ac o s, and (ii) masked s imuli
iden i ica ion, wi h block (B1 s. B3), condi ion (Mask22 and Mask33), and s imuli ype
(Bio s. Sc ) as ac o s.
The mANOVA esul s showed no signi ican main e ec s o block o any in e ac ions
wi h o he ac o s. Consequen ly, we analysed all ials om all h ee expe imen al blocks
oge he . Fo all es s, we adop ed alpha = 5%.
The accu acy da a showed a no mal dis ibu ion acco ding o he Shapi o–Wilk no -
mali y es . The mANOVA o accu acy, conside ing condi ion (NoMask s. Mask22
s. Mask33) and s imuli (Bio s. Sc ), e ealed a main e ec o s imuli (F1,11 = 8.29,
Neu oSci 2025,6, 2 7 o 11
p= 0.015
,
η
p2 = 0.430, BF10 = 43.51), which showed highe accu acy o biological e sus
sc ambled PLDs, and o condi ion (F2,22 = 33.32, p< 0.001,
η
p2 = 0.752, BF10 = 873.34)
and s imuli*condi ion in e ac ion (F2,22 = 11.43, p< 0.001,
η
p2 = 0.510, BF10 = 335,484.32).
Bon e oni-co ec ed pos hoc analysis e ealed highe accu acy o NoMask compa ed
o Mask22 (p< 0.001, BF10 = 545.897) and Mask33 (p< 0.001, BF10 = 2264.784) bu no
di e ences be ween Mask22 and Mask33 (p= 0.777, BF10 = 0.339). Bon e oni-co ec ed
pos hoc o he in e ac ion showed ha accu a e esponses o he biological PLDs in he
NoMask condi ion we e highe han o all o he s imuli: BioMask22 (p= 0.004), BioMask33
(
p< 0.001
), Sc NoMask (p= 0.003), Sc Mask22 (p< 0.001), and Sc Mask33 (p< 0.001). S a-
is ical esul s a e illus a ed in Figu e 3, and he mean accu acy o each s imuli ype is
depic ed in Table 1.
Neu oSci 2025, 6, x FOR PEER REVIEW 7 o 11
The mANOVA esul s showed no signi ican main effec s o block o any in e ac-
ions wi h o he ac o s. Consequen ly, we analysed all ials om all h ee expe imen al
blocks oge he . Fo all es s, we adop ed alpha = 5%.
The accu acy da a showed a no mal dis ibu ion acco ding o he Shapi o–Wilk no -
mali y es . The mANOVA o accu acy, conside ing condi ion (NoMask s. Mask22 s.
Mask33) and s imuli (Bio s. Sc ), e ealed a main effec o s imuli (F1,11 = 8.29, p = 0.015,
ηp2 = 0.430, BF10 = 43.51), which showed highe accu acy o biological e sus sc ambled
PLDs, and o condi ion (F2,22 = 33.32, p < 0.001, ηp2 = 0.752, BF10 = 873.34) and s imuli*con-
di ion in e ac ion (F2,22 = 11.43, p < 0.001, ηp2 = 0.510, BF10 = 335484.32). Bon e oni-co -
ec ed pos hoc analysis e ealed highe accu acy o NoMask compa ed o Mask22 (p <
0.001, BF10 = 545.897) and Mask33 (p < 0.001, BF10 = 2264.784) bu no diffe ences be ween
Mask22 and Mask33 (p = 0.777, BF10 = 0.339). Bon e oni-co ec ed pos hoc o he in e -
ac ion showed ha accu a e esponses o he biological PLDs in he NoMask condi ion
we e highe han o all o he s imuli: BioMask22 (p = 0.004), BioMask33 (p < 0.001), Sc No-
Mask (p = 0.003), Sc Mask22 (p < 0.001), and Sc Mask33 (p < 0.001). S a is ical esul s a e
illus a ed in Figu e 3, and he mean accu acy o each s imuli ype is depic ed in Table 1.
Table 1. Mean numbe o ideos wi h co ec esponse; s anda d e o and con idence in e al o
each s imuli ype.
Condi ion
S imuli
Mean
SE
95% CI (Lowe /Uppe )
NoMask
Biological
12.50
0.702
10.96/14.04
NoMask
Sc amble
6.17
1.347
3.20/9.13
Mask22
Biological
6.58
1.246
3.84/9.33
Mask22
Sc amble
4.33
0.980
2.18/6.49
Mask33
Biological
5.08
0.743
3.45/6.72
Mask33
Sc amble
4.42
0.857
2.53/6.30
Figu e 3. Panel (A) shows means and con idence in e als o accu acy acco ding o condi ion. As
can be seen, accu acy o NoMask (9.3 ± 0.897 SE) > Mask22 (5.46 ± 0.849 SE) = Mask33 (4.75 ± 0.542
SE). Panel (B) shows means and con idence in e als o accu acy acco ding o s imuli and condi ion.
As can be seen in his panel, accu acy o biological (8.06 ± 0.735) > sc ambled (4.97 ± 1.001) s imuli.
Fu he , he accu acy o BioNoMask s imuli was signi ican ly highe han o all o he condi ions.
* indica es p < 0.001, and ** indica es p < 0.05.
Fo he eac ion ime analysis, we included only ials in which pa icipan s’ e-
sponses we e co ec . Acco ding o he Shapi o–Wilk no mali y es , all condi ions excep
o Sc NoMask (p = 0.033) showed a no mal dis ibu ion. Consequen ly, p- alues a e
Figu e 3. Panel (A) shows means and con idence in e als o accu acy acco ding o condi ion. As can
be seen, accu acy o NoMask (9.3
±
0.897 SE) > Mask22 (5.46
±
0.849 SE) = Mask33 (4.75
±
0.542 SE).
Panel (B) shows means and con idence in e als o accu acy acco ding o s imuli and condi ion. As
can be seen in his panel, accu acy o biological (8.06
±
0.735) > sc ambled (4.97
±
1.001) s imuli.
Fu he , he accu acy o BioNoMask s imuli was signi ican ly highe han o all o he condi ions.
* indica es p< 0.001, and ** indica es p< 0.05.
Table 1. Mean numbe o ideos wi h co ec esponse; s anda d e o and con idence in e al o
each s imuli ype.
Condi ion S imuli Mean SE 95% CI (Lowe /Uppe )
NoMask Biological 12.50 0.702 10.96/14.04
NoMask Sc amble 6.17 1.347 3.20/9.13
Mask22 Biological 6.58 1.246 3.84/9.33
Mask22 Sc amble 4.33 0.980 2.18/6.49
Mask33 Biological 5.08 0.743 3.45/6.72
Mask33 Sc amble 4.42 0.857 2.53/6.30
Fo he eac ion ime analysis, we included only ials in which pa icipan s’ e-
sponses we e co ec . Acco ding o he Shapi o–Wilk no mali y es , all condi ions excep
o Sc NoMask (p= 0.033) showed a no mal dis ibu ion. Consequen ly, p- alues a e
epo ed wi h he G eenhouse–Geisse co ec ion applied. The mANOVA o eac ion
ime, conside ing condi ion (NoMask s. Mask22 s. Mask33) and s imuli (Bio s. Sc ),
showed no main e ec o s imuli (F1,8 = 0.134, p= 0.724,
η
p2 = 0.016, BF10 = 0.270), condi-
ion (
F2,16 = 2.520
,p = 0.116,
η
p2 = 0.240, BF10 = 0.249), o s imuli*condi ion in e ac ion
(F2,16 = 1.003, p= 0.377, ηp2 = 0.114, BF10 = 0.071).
Neu oSci 2025,6, 2 8 o 11
To u he in es iga e he e icacy o he masks in iden i ying a human agen in
each o he biological ac ions, we summed he o al numbe o co ec answe s o each
mo emen ac oss he 12 pa icipan s and calcula ed he pe cen age accu acy using he
o mula Nca/Ts im
×
100 (whe e Nca = numbe o co ec answe s, and Ts im = o al
numbe o he speci ic biological s imuli in each condi ion, which is 24 o each ac ion).
As shown in Table 2, he iden i ica ion o s imuli dec eased wi h bo h masks o nea ly
all mo emen s. In e es ingly, o he Jump Rope ac ion, he mask con aining 22 do s
educed he accu acy o 25%, while accu acy emained he same o he NoMask and
Mask33 condi ions.
Table 2. Pe cen age accu acy in de ec ing a human ac ion in each biological s imulus acco ding
o condi ion.
Biological Mo ion NoMask (%) Mask22 (%) Mask33 (%)
Jumping Jack 95.83 45.83 45.83
Walk 70.83 66.66 45.83
La e al S ep 87.50 58.33 50
Ma ch 91.67 50 50
Low Kick 95.83 37.50 20.83
Jump 91.67 45.83 29.17
Jump Rope 91.67 25 91.67
4. Discussion and Conclusions
This a icle p oposes and explains a use - iendly app oach o c ea ing isual noise
masks o PLDs. We chose Blende so wa e [
21
], which allows he cons uc ion and
modi ica ion o ideos. The so wa e seems aluable o esea che s s udying isual
pe cep ion o anima ed ac ions, as well as o p o essional anima o s. I enables he c ea ion
o s imuli om sc a ch o he edi ing o p e iously eco ded mo ion da a and hei masks.
Ano he excep ional ea u e o his p og am is he abili y o wo k wi h bo h 2D and 3D
o ma s, o e ing buil -in ideo edi ing ools. Fu he mo e, he so wa e’s lexibili y in
manipula ing bo h 2D and 3D ile o ma s makes i pa icula ly use ul o edi ing body
kinema ics eco ded wi h mo ion cap u e sys ems like Vicon [
25
,
26
] and o adap ing noise
masks acco dingly. Al hough he so wa e may encoun e limi a ions wi h high- esolu ion
anima ions o complex physics simula ions, as ende ing imes can inc ease signi ican ly in
mo e demanding p ojec s [
27
], i s e iciency in gene a ing simple isual s imuli makes i an
excellen ool o ou pu pose.
We success ully buil ou noise mask o PLDs selec ed om a alida ed da abase [
4
]
and desc ibed he me hods o guide esea che s in p oducing simila s imuli. Fu he mo e,
we es ed he e icacy o he masks in a simple o ced-choice ask, whe e pa icipan s had
o indica e whe he hey could see a human ac ion in bo h he unmasked and masked
ideos. Reac ion ime esul s e ealed no signi ican di e ences be ween condi ions, likely
because ou s imuli we e p esen ed o only 1 s, and esponses we e only egis e ed
i p o ided wi hin his ime ame, he eby inc easing he ask’s di icul y. Indeed, he
accu acy esul s e ealed a s a is ically signi ican main e ec o condi ion, showing
ha pa icipan s made mo e mis akes when he PLDs we e p esen ed wi h dynamic 22-
and 33-do masks. E o s included ailing o de ec human ac ions in biological mo ion,
inco ec ly iden i ying ac ions in sc ambled mo ion, o no esponding in ime. This
p o ides empi ical e idence o he e ec i eness o masking. No ably, when e alua ing
he pe cen age accu acy o each biological s imulus, he e was a gene al dec ease when
p esen ed wi h masks, excep o he Jumping Rope s imulus in he Mask33 condi ion. I is
impo an o emphasise ha he expe imen was designed o e alua e he masks’ e ec s a
he condi ion le el (i.e., NoMask, Mask22, and Mask33), as is ypical in ac ion pe cep ion
Neu oSci 2025,6, 2 9 o 11
expe imen s [
6
,
10
,
22
,
23
]. In es iga ing hei e ec s on speci ic ac ions would equi e
signi ican ly mo e epe i ions o each ac ion ype. None heless, bo h masks demons a ed
a obus gene al e ec in educing ac ion iden i ica ion, wi h s a is ically signi ican esul s
and la ge e ec sizes.
The indings sugges ha he 22-do mask is e ec i e o his da ase and o he
simila ones. Fu he mo e, as he e was no signi ican di e ence be ween he 22-do and
33-do masks, we conside he 22-do mask e ec i e o his da ase and o he simila
ones. While mo e complex masks wi h a ying do s and mo ion ajec o ies could be an
in e es ing di ec ion o u u e esea ch, designs wi h oo many do s may o e ly obscu e
s imuli, making hem i ually un ecognisable. Addi ionally, as s anda disa ion wi hin he
s udy is essen ial, we op ed o c ea e a single mo ion ajec o y, ein o cing he p oposed
app oach’s p ac icali y.
We ha e made ou masks and he Blende p ojec a ailable o acili a e he wo k o
in es iga o s who need o use hese p ocedu es. In ou Blende p ojec , i is also possible
o se a di e en amoun o ames pe second o each o he do s o depic hei ajec o y
and speed, e en hough we did no use ha ea u e because ou goal was o ha e he same
smoo hness and speed o all do s. Addi ionally, i is possible o edi ou p ojec by copying
o dele ing he cu en do s o de ine he desi ed numbe o u u e expe imen s.
This edi ing lexibili y is impo an because he con igu a ion used he ein was speci i-
cally ailo ed o he ideos and da abase se ing as he basis o his expe imen . Thus, he
pa ame e s used in ou expe imen should no be conside ed uni e sal o all ypes o PLD
s imuli. I is impo an o emphasise ha ou p ima y goal was o p esen an al e na i e
o p og amming o c ea ing PLD noise a he han o de elop a mask ha i s all s imuli.
While ou app oach is gene alisable, we ecommend adjus ing he pa ame e s as needed o
di e en s imuli and esea ch objec i es. In summa y, while he alues ob ained canno be
di ec ly applied o o he con ex s wi hou i s unde going an e alua ion, he me hodology
desc ibed in he manusc ip o de e mine he pa ame e s is eplicable.
In his con ex , he in en ion o his wo k was no o alida e hese speci ic masks
o se e al da ase s bu a he o p o ide a s aigh o wa d and use - iendly me hod o
c ea ing and applying masks o PLDs. The main ad an ages o he p esen ed me hodology
a e i s cos -e ec i eness and he lack o need o p og amming expe ience, making i
accessible wi hou any p e equisi es. This s ep-by-s ep guide may pa icula ly appeal o
s uden s engaged in his esea ch opic who a e no ices in he p og amming skills ypically
equi ed o c ea e isual do noise masks, such as algo i hms [
9
], o e ing a aluable
esou ce o explo e mo ion pe cep ion using PLDs.
While we ha e aimed o make his me hodology accessible and use - iendly, we
ecognise ha any new app oach, ega dless o i s simplici y, ca ies he isk o ailu e o
limi a ions. The e ec i eness o his me hodology can be e alua ed h ough ca e ul imple-
men a ion and es ing in each speci ic con ex . As such, we ecommend ha esea che s
alida e and adap he pa ame e s as needed o hei own da ase s and esea ch goals,
ensu ing he necessa y expe imen al igou is main ained.
Finally, we encou age o he esea che s and anima o s o explo e his and simila
pla o ms u he o expand he scope o he p o ided guidelines. These guidelines a e
cu en ly limi ed o a speci ic ype o s imuli, as we ocused on human ac ion pe cep ion,
bu hey may ha e b oade applicabili y. Fu u e esea ch could examine hese masks in
domains beyond human mo ion. Fo example, his app oach could be adap ed o s udying
human analysis o animal mo ion, such as gai o ligh pa e ns, as well as o expe imen s
es ing whe he animals pe cei e biological mo ion simila ly o humans [
28
]. Addi ionally,
hese masks could be use ul o objec iden i ica ion s udies o explo e how humans o AI
sys ems de ec mo ing objec s unde isually noisy condi ions. The lexibili y o he masks,
